On a bright March afternoon, a slender, 50ft-long boat made her maiden voyage across calm Cornish waters, powered by the paddle-strokes of an 18-strong crew. With a hull made from oak planks shaped using bronze-bladed tools and painstakingly stitched together with flexible twists of yew, the vessel represented the very latest in boat-building technology in c.2000 BC – while the launch took place in AD 2013.
Named after a sea monster reputed to lurk beneath the surface of Falmouth Bay, Morgawr was built at the National Maritime Museum Cornwall (NMMC) as part of an archaeological experiment, led by Professor Robert Van de Noort (then of the University of Exeter, now at the University of Reading), to investigate how sewn-plank boats were built and used during the Early Bronze Age.
These vessels marked a technological step-change from the dug-out canoes and skin boats used by earlier communities, but as their development pre-dates the widespread use of iron by several centuries, their hulls were sewn, rather than nailed, together. England and Wales boast the oldest known craft of this type in Atlantic Europe; the remains of ten have been discovered so far, ranging in date from c.2030 to c.760 BC. While portions of their hulls have been preserved in estuary mud and waterlogged sites for centuries, these remains are all tantalisingly fragmentary – leading to debates about what a complete boat might have looked like, and whether it was capable of operating on the open sea.
It was to investigate such questions that, in March 2012, planning began for the building of a full-size replica, drawing together all available archaeological information and using only materials, tools, and techniques available during the Bronze Age (and funded by a generous Arts and Humanities Research Council grant).
The chief source of information for Morgawr’s appearance came from a series of three sewn boats that were discovered on the Humber foreshore between 1937 and 1963. Regrettably, neither Ferriby 1 (c.1880-1680 BC) nor its sister boat, Ferriby 2 (c.1940-1720 BC), have survived to enable direct analysis – the remains had been discarded after early efforts to conserve them failed. The oldest of the three craft, Ferriby 3 (c.2030-1780 BC), was spared this fate, and is now housed in the Hull and East Yorkshire Museum, but this adds a limited amount to our understanding of how a complete boat looked, because only a single outer bottom plank, connected to a fragment of the lowest side strake, has survived from this vessel.
All is not lost, however: the excavation of Ferriby 1 and 2 – undertaken in 1946 by the brothers Ted and Willy Wright, with the help of C W Phillips, who also worked on the Sutton Hoo dig – was carefully recorded, leaving us with an invaluable archive of field sketches, photographs, and measurements from which to work. From drawings of Ferriby 1 as it appeared in situ – its two ends exposed and the depth of the keel plank established with the help of a pointed stick – we can tell that the craft’s lengthwise profile was curved, with the ends rising 1.32m higher than the centre of its keel.
Previous attempts at reconstructing the appearance of Ferriby 1 gave it a flat punt-like base. Examination of the field archives and the remains of Ferriby 3 established that the keels of these boats in fact formed a very shallow V-shape: the misleading flat appearance had been due to the slumping of the waterlogged timbers when placed on a hard floor post-excavation.
As early as 1989, such details had already been fed into a reconstruction blueprint, dubbed the ‘complete boat’, by Ted Wright (the discoverer of Ferriby 1) and John Coates (a naval architect specialising in ancient ships). This incorporated evidence from Ferriby 1, 2, and 3, together with a certain degree of educated guesswork, as no part of the upper hull of any of the trio has survived. This model proposes a craft measuring 15.9m in length, with a maximum beam of 2.52m. Its sides rise three strakes high, with thwarts (cross-planks) to improve the structural integrity and provide seats for the boat’s 18 paddlers.
This design was adapted slightly by the Morgawr team for practical reasons, reducing its dimensions slightly to 15.28m by 2m, creating finer lines that allowed for more efficient paddling. Robert explained that: ‘the Wright and Coates reconstruction asks for a wider boat than Morgawr, but when we tried laying that out it just wouldn’t work. We had people in different quarters paddling at quite different angles – they were fighting each other for direction, and wasting a lot of energy. When we reduced the width to 2m, it worked much better.’
A floating hypothesis
Half-sized reconstructions of the Bronze Age sewn boats had been built in the past, with the Ole Crumlin-Pedersen, a scaled-down version of the Dover Boat, first launched by Canterbury Archaeological Trust in 2012 (CA 287), and before that Joyce and Edwin Gifford had created the Oakleaf, a Ferriby-type vessel also drawing on Wright and Coates’ ‘complete boat’ diagram (CA 191).
The reasons for building a scaled-down model are understandable – reducing your boat to half size also significantly reduces production, maintenance, crewing, and storage costs – but it also detracts from the boat’s ability to tell us how the real thing might have handled, and they create misleading perceptions of how these vessels looked – particularly when they are crewed by full-sized humans. And, Robert says, ‘the practical characteristics of the materials themselves, such as tension and flexibility, cannot be scaled up or down, which makes testing them more difficult’. Thus it was decided that Morgawr would be built full-sized.
As well as reconstructing the boat, the team planned to use replicas of Bronze Age tools. Two distinct toolmarks have been identified on the surface of the Ferriby boats, but these have been overcut, partly obscuring them and making exact reproduction difficult, so additional evidence was sought from excavated examples of flat-bladed axes, adzes, and chisels from Yorkshire. This led to the creation by Neil Burridge, an expert bronze-worker who had previously run a number of experimental programmes with the University of Exeter’s archaeology department, of a toolkit comprising 27 replicas made from 88% copper and 12% tin, similar to the composition of their prehistoric predecessors. Neil cast each tool in an individual stone mould, before quenching with cold water and hammering the blades for around ten minutes. They were then polished and sharpened, before the blades were fitted to ash handles with leather lashings.
Spearheading the construction process itself was Brian Cumby, a shipwright brought on board for his more than 30 years’ experience in building wooden craft, including a number of reconstruction projects. Overseen by archaeologists from the universities of Southampton and Exeter and NMMC staff, he worked with a team of 100 volunteers, ranging in age from teenagers to octogenarians. These came from all walks of life, including archaeology and fine-art students, as well as retired boat-builders, engineers, management consultants, and sailors. After three months’ training in the use of Bronze Age tools and the design of sewn-plank boats at the University of Exeter, he was ready to begin.
The arrival of 20 tonnes of timber, cut from three English oaks, heralded the start of the building phase of the project in April 2012. The high cost of timber meant that these oaks could not be rejected if they turned out to be less than ideally suited to the project – an economic handicap not experienced by prehistoric shipwrights. So, keen as the volunteers were to get to work, and determined as everyone was only to use materials and techniques that would have been available in the Bronze Age, the importance of getting the first step right meant that some compromises had to be made. After much deliberation, it was decided not to split one of the logs to craft the keel plank from the two halves: there was too much risk of not producing two matching halves, and the budget did not allow for a second attempt, so the team agreed to have the keel plank’s tree sawn in two.
This done, work began to debark and roughly square the trunks using flat-bladed axes, before adzes were employed for the finer shaping. Here, Robert said, an unexpected new aspect of the project arose. ‘We discovered that the young male volunteers found using the replica tools much more difficult than the women on the team,’ he said. ‘This was because they were going at it too hard, and just whacking the timber; our female volunteers worked more delicately. The boys seemed to prize strength over technique, and a few hafts got broken in the process, revealing an interesting sociological dimension to the experiment.’
Working the timbers by hand was a lengthy process, but the keel plank was complete by the end of July, the inner cleats had been chiselled into shape by early September, and a month later the team finished stitching the outer bottom planks. Thin strips of yew (withies), produced by twisting branches to create a split that could then be worked along their whole length, proved to be a very effective material for this purpose. The tensile qualities that have made it a popular choice for making bows throughout history made yew ideally suited to sewing: it is also more durable than other flexible woods, like willow and hazel, and rather easier to harvest and handle than holly.
Stitching the planks together
The original plan had been to use withies donated six months earlier by the Westonbirt National Arboretum, but the stems had lost their sap and completely dried out by the time the team came to use them: no amount of soaking could restore their flexibility – another useful lesson in prehistoric wood-use learned by accident. Local graveyards came to the rescue with a supply of fresh material, which the volunteers harvested no more than two days in advance. They also discovered that storing the withies submerged in water helped to prolong their working life.
Countersunk into the bottom of the hull to protect them during beaching, some 225 stitches were created in all, threaded through 450 holes. These had to be chiselled in two directions, before the shape was carefully rounded to get rid of any sharp corners or rough edges that could damage the stitches. Each hole took an hour to fashion – a vital but time-consuming process.
As each stitch was completed, the withy end was threaded back underneath and trimmed. Gaps were plugged with a wad of moss, mixed with tallow to prevent the caulking from drying out, and a layer of beeswax was spread over the top for waterproofing. These final steps represent another small compromise by the team, as there is no archaeological evidence for tallow or beeswax being used in any Bronze Age boat found so far. Both materials would have been readily available during this period, however, and if they had been involved in the prehistoric caulking process, it is likely that any trace would since have been washed away.
Strakes and strokes
Three side strakes completed the height of the hull, painstakingly chipped into shape by adze-wielding volunteers. Initially they had used wooden wedges to help remove excess timber, but this proved to be a false economy, as the resulting uneven surfaces were difficult to work with bronze tools. The team had also ruled out steaming the planks to help bend them, a technique used in other reconstructions such as the Oakleaf, since Bronze Age technologies like open fires and hot rocks would not have produced enough heat to affect Morgawr’s thicker timbers.
Shaping the outer surface of the planks turned out to be a fairly straightforward (if lengthy) task, though the process of fitting them raised further questions about prehistoric boat-building: were template frames used to help guide the shape of a hull, or were Bronze Age shipwrights, being more accustomed to working with sewn-plank technology, able to simply piece vessels together ‘by eye’? Construction aids are largely invisible in the archaeological record – though, if they had been used, they would have been removed after the boat was completed – but the team did find it useful to use temporary patterns to stitch the planks and maintain a more uniform shape.
A final question concerned the matter of propulsion. Happily, suitable paddles were readily available in the archaeological record: one was found near Ferriby 1 in 1939, and a second was recovered from Ferriby 2 seven years later. Using these as a guide, the Morgawr volunteers were encouraged to craft paddles with handle lengths that suited them individually. The discovery of paddles of various lengths with the Danish Hjortspring boat (c.350 BC) suggests that this could have been a plausible prehistoric practice.
Moment of truth
The resulting paddles were put to work on 6 March 2013, almost a year after construction began, as Morgawr was launched into the cold waters of Falmouth harbour, crewed by some of the volunteers who built her, with Robert at the helm. As most new boats do, before their soaked timbers swell and close any remaining gaps, she immediately took on large amounts of water. Her crew were neither surprised nor alarmed by this, however, and after an energetic burst of bailing the boat was able to complete two 500m trial-runs before returning triumphantly to dry land.
Since then, Morgawr has made a number of longer outings around the harbour, using a crew of experienced paddlers from the Helford River Gig Club; plans are now being formulated for a series of more ambitious sea-trials, taking the craft further out to see how she behaves under different weather conditions.
One question the team hopes to answer through such trials is whether a boat as long and thin as Morgawr could travel safely at sea, or whether sewn boats were limited to river journeys. Unlike Bronze Age logboats, which are more commonly found at the bottom of rivers (CA 263) and lakes (CA 292) – sewn boats are typically found in estuaries and coastal sites, suggesting that their scope was not limited to inland travel. At the very least, it seems likely that they were capable of making longer journeys by hugging the coast, something suggested by the discovery of Kimmeridge shale from the Dorset coast within the hull of a sewn-plank boat found in Dover, 120 nautical miles away (CA 133).
There are other clues that point to more ambitious Bronze Age voyages: artefacts from this period point to a thriving network of trade contacts with Continental Europe, particularly in metal objects. The Moor Sand wreck (CA 243) may well have been heading across the Channel to the Continent when it was lost off Salcombe 3,000 years ago with its cargo of copper and tin ingots, as well as metal weapons, tools, and ornaments.
Such trade is unsurprising given the mineral wealth of the south-west, particularly Cornwall, which supplied much of Roman Europe with tin. Recent analysis of the remarkable lunar images adorning Germany’s Nebra Sky Disk suggest that its inlays were made from Cornish tin and gold. On this evidence, Bronze Age vessels were certainly capable of making sea voyages, and Robert reports that Morgawr has proven to be a reliable boat – if awkward to handle – once the members of the team became accustomed to paddling in unison and maintaining a straight course with the wind across the bow.
‘Weighing 7 tonnes when wet, she is very heavy and unwieldy – it is hard work to get her started, and if you stop paddling she slows down very quickly – but it is quite pleasant once you get going, and steering got easier as we became more experienced at paddling,’ he said. ‘In any case, there would have been fewer requirements for a Bronze Age boat to be steered accurately, as harbours would have been natural beaches and salt marshes, and tidal movement would have been used to navigate between land and sea without needing jetties. Despite being long and thin, Morgawr is remarkably stable, easily going through waves a metre high.’
The Bronze Age techniques used in the construction of Morgawr have also stood the team in good stead: 18 months after her launch, the boat’s timbers show no sign of deterioration, and every yew stitch remains intact. Her cleats and transverse timbers also proved effective at securing the hull, with only minimal movement of the keel and outer bottom planks discernible while the craft was in motion. This should not be the final word in reconstructing vessels like these, though, Robert insists. ‘The next thing we would like to see is other groups challenging us and proposing alternative models of how these vessels might have looked,’ he said. ‘This is the best way of going forward in academic discussions.’
Further reading: Robert Van de Noort et al., ‘Morgawr: an experimental Bronze Age-type sewn-plank craft based on the Ferriby boats’, The International Journal of Nautical Archaeology (2014) 43.2
ALL IMAGES: Robert Van de Noort, unless otherwise stated.