Digging Deeper



The Great Rudder Mystery

Supporting the rudder

When the ship was excavated it was found to have two closely spaced frames or ribs at the stern. Frames 24 and 25 seem to indicate the hull had strengthening needed to support … something? It does not take a great leap of the imagination to suggest a solution. They must have supported the rudder. That is obvious - but how? During the 1939 excavations the investigators searched this area for the next part of the evidence, by digging on the outside of the hull to find the rudder boss, or button, or a strip of timber. This extra piece of wood supports the planking at a pivot point of the rudder spreading out the load that it will impose on the hull. The boss also extends outwards. 

As the hull at this point is ‘V’ shaped a vertical rudder would touch the gunwale at the top, but then be unsupported below it. It would tend to be pushed back against the hull, or be pulled outwards by the force of the water. It would be waving back and forth, unless it was secured in a second place lower down. (See notes on the Nydam ship later.) The outer tip of the boss can be in line with the gunwale above it, so that it provides a second support point. The rudder can be attached to the hull by a tight lashing through a hole in the boss that will allow the rotating movement needed to steer the ship. This is the basic reasoning behind looking for the boss. If it could be found, it would confirm the usual expectations relating to the rudder. No evidence of anything was found. This then posed a big question about how the rudder could be attached to the ship.

Volume 1. Section 7. Page 406. The steering system

                There are several photographs taken by Phillips, Lack and Wagstaff in 1939 and probably the best one is by Phillips, page 406. There are descriptions of the related fastenings and their sizes as well as the dimensions of the ribs as precisely as they could be determined by looking at the evidence and examining the rivets. There is a diagram and a brief written description by Phillips. The wider part of rib 24 (see diagram above) is secured by 5 rivets so he used the word ‘quincunx’ to describe that situation, though he slightly confused the situation by not using the correct number for describing the rib involved. A highly accurate, dimensioned drawing of this part of rib 24 is Detail J, top left corner of Hutchison’s drawing, Science Museum 2012b in the Ipswich Museum.

                This section of the ship impression was destroyed during the war, so it was not possible to investigate further. As with almost everything about the ship there are many ‘what ifs’. For example, did Hutchison look far enough back on the hull for the button? Many steering systems from the Scandinavian ships have rudders raked aft, so the button would not be in line with frame 24, but astern of it. We cannot know, but we can assume that he searched as well as circumstances allowed.

We are left with a bit of a mystery. How was the rudder attached?


The evidence

The question has not gone away, because any reconstruction must furnish the ship with a functioning rudder. In the many years since the excavations it has not been a pressing problem; it could conveniently be overlooked, but that is now no longer the case. A steering system based on the evidence of the original must be found. This might seem an insurmountable problem, but it must be addressed.

When any functioning system is designed or evolved, the people making it, look at each part and shape it according to the way it works within the system. When we look at frames 24 & 25, that are set 18ins apart, we see that:

  • they are made much wider with extra fastenings at the top than lower down. This suggests that the top, near the gunwale was where the load or wear was imposed. If there had been a boss-type system there may have been strengthening lower down instead of higher up
  • The front frame 24 is more heavily strengthened than the rear one 25. This probably implies that the rudder support point is somewhere close to this frame. If this was the case, the forward frame would take the majority of the load. Water pushing the rudder backwards would generate the highest load. It could be that a support bracket back to frame 25 would help resist this load

Before investigating further possibilities it is worth considering the wider picture concerning the influences on ship construction around and well before the year 600AD.

The external appearance of the Sutton Hoo ship delivers an immediate ‘Scandinavian’ message. There has been a tendency to compartmentalise historical study into blocks, Roman, Anglo-Saxon, Viking. Today there is a greater awareness that a broad blend of influences could amalgamate to produce an outcome. The Roman and Anglo-Saxon ‘periods’ did not have a clinically sharp divide in 410AD. In all probability, in shipbuilding, the two traditions influenced each other for several hundreds of years in an amalgamated continuum.


The Roman digression

The area of Britain where this ship was discovered had been under the influence of the Romans for about 400 years until a couple of hundred years before the Sutton Hoo ship was built. The Romans built, used and maintained ships. They lived in this location, they used the rivers for communication and transport in just the same way as the Anglo-Saxons who overlapped with them and followed after them. There would have been an exchange of knowledge and influence between the two groups.

The ship that most closely resembles the Sutton Hoo ship is Nydam and it is interesting to read their understanding of cross-culture influences in The Nydam ship (Das Nydambåden)  by Angelica Abegg-Wigg. ‘Different features in the construction and equipment of the boat from Nydam confirm provincial Roman influences. Boat builders from the Iron Age in Northern Europe adopted this technology and modified it. Both lapstrake technique and caulking with textile materials were known in the Mediterranean Region and in ancient Gaul. The nailed scarf joint bonding the ends of the keel with the stem and stern copies a provincial technique. Furthermore the technique of inserting massive ribs into the raw hull can be seen in every Roman planked ship.

The Mainz ships (see below) are described as:

‘designed for rowing and fitted with sails. They were possibly used for patrolling, as courier boats and for transporting troops.’

Of the Sutton Hoo ship it says:

‘It was a boat designed for rowing. There are no indications of sail ever having been used.’


The Nydam rudders are of special importance and will be mentioned later.

It is known that German seafarers were clashing with Roman craft in the North Sea area from the first century BC. The information can be found in Dark Age Naval Power by John Haywood. Haywood also explains that it is known that an auxiliary cohort of men from the middle Rhine area serving under Agricola on the west coast of Britain seized three Liburnian galleys and sought to return them to the Rhine region by rounding the coast of northern Britain. Their story can be read in the book, but our concern is simply with the fact that men such as these were working on similar ships, possibly for the whole period of Roman dominance in the region. In 500 years of contact there must have been a flow of information between them. People adopt better, useful or time-saving ideas.

The Museum of Ancient Shipping in Mainz, Germany has reconstructions of some craft discovered there during the extension of a hotel in the 1980s. Close studies have been made of 5 of the 10 craft and a couple of the craft known as Type Mainz A are of particular interest. The latest craft is dated from the end of the 4th century. One is 21.6m long and had 30 oars. It also had sailing equipment, but it is categorised as a small troop transporter or rapid rowing ship. It is carvel built with planks arranged edge to edge in the Mediterranean style. The planks are attached to rectangular ribs 7cms x 12cms. It is a Roman military ship, locally built with German oak, as are all the craft.


Although the Sutton Hoo ship was constructed 200 years later than the Mainz ships they have in common that they all used rectangular ribs. This is a very important point to note. True Scandinavian vessels of this size might have been using inverted pear-shaped ribs tied to cleats. Nydam 300 years before Sutton Hoo used tied planks, Kvalsund roughly contemporary with Sutton Hoo used them and Oseberg 400 years after Sutton Hoo used them. The use of plain rectangular frames in the Sutton Hoo ship implies a possible Roman influence. The exterior of the ship with its clinker planking and upswept ends is unmistakeably Scandinavian, but its ribs seem to indicate that it borrowed from Mediterranean-style ribs and this could indicate that it was built in Britain where the Roman influence would have been strong. Its function was similar to the Mainz ships, to rapidly move a body of men irrespective of wind direction, in shallow rivers. The Sutton Hoo ship had the added requirement of riding modest waves at sea that the Mainz ships, being 250 miles inland, would not usually be required to do. If we choose to accept these factors we can see that a blend of influences may have resulted in the Sutton Hoo ship’s form and construction. It might therefore, be helpful to look at the rudder arrangements on the Roman ships. They could have influenced the Sutton Hoo ship, just as the Roman ribs may have done.


A second look at the Sutton Hoo ribs 24 & 25

  • If we wished to make the frames wider at the top, the natural way would be to just increase their width evenly on both sides of the frame – either an inverted triangle or a kind of frying pan shape. This is not what has been done. This is not a casual accident. The ribs are clearly deliberately widened on one side only. This creates an additional angled edge below the forward face. Looking at the photographs of the excavation all the edges of the ribs have lost their original sharpness. Everything is rounded to some extent, but originally this extra downward facing surface could have had a function. On all of the drawings the expanded form of the top of the ribs is represented with straight, sharply defined edges.
  • Ribs aft of about No 20 are set at about 90° to the gunwale, not the keel, except the last frame 26. This is not an accident; it was intended. Why?
  • The widening suggests a thicker support surface for a downward force from above

It is by observing such details as this that it becomes possible to try to learn more of what was in the builder’s mind when the decisions were first taken. Is frame 26 upright as it formed part of a level platform for the helm-person? If this is the case, should frame 25 also be upright to support the platform at its front edge? The answer is No. It was more important for frame 25 to be part of the rudder support system than to be part of the platform.


The Roman rudder system

The illustrations are taken from the Museum of Ancient Shipping, Mainz website. They may not all refer to one boat. The craft all had some features in common. The wrecks were incomplete or damaged by current building work, so it was necessary to translate some parts from one craft to another. It will be seen that the long-buried timber is not pristine and square edged as has been mentioned about the hardened sand impressions of the Sutton Hoo ship.

Left. This is part of the strengthening rudder-support structure of a ship, as found.


Right. Is an interpretation as made for a reconstruction.


This is a model of another excavated ship

called Oberstimm that dated from about 100AD. The shape of the bottom edge of the rudder blade should be noted, as it relates to the way water behaves – hydrodynamics. It is not likely that the Sutton Hoo ship used this adaption which functions best at sailing speed.


It will be seen that the Romans used two rudders, one on either side of the ship. They were fixed to a beam that ran right across the ship, piercing through the top plank just below the gunwale. The beam has additional support from shaped diagonal brackets bracing it to the next frames astern. Looking at the stress/ loadings involved, as the rudder stock is moved further outwards, away from the gunwale, the leverage will increase on the support beam. A rearward support frame will then be required. It will be pushed inwards against the hull by the rudder stock loading that is positioned ahead of it.

        This Is a reconstruction of a Mainz craft showing the details. There was probably no actual component part discovered for the timber illustrated as being forward of the support beam which is represented as a different material. There are no lashings visible. The rudders would be free to move up and down as they are, but would lock with forward movement in water. There are extra support brackets behind the rudder support beam.


A rudder system for the Sutton Hoo ship

Are you ready to take the challenge of designing a functioning system for the Sutton Hoo ship?

The current situation is that:

  • there are 2 shaped ribs or frames. We don’t know how accurately the representations are of what was found, nor how thorough was the search for additional detail
  • it is possible that there were holes or trenails that were never discovered
  • the rudder support system could be removed from the ship
  • it was thought to be sufficiently valuable to remove the rudder assembly, but all internal components such as the seats (thwarts) were also removed

There are some additional possibilities that arise from looking not only at Scandinavian rudder solutions, but at Roman ones also. If the upper parts of ribs 24 & 25 had been the same shape on the opposite side of the ship it would have been a clear indication that something akin to the Roman system was intended. The ribs were not recorded as being shaped. If a rudder support beam could sit on top of the gunwale it would open up a number of possibilities. The Sutton Hoo diagram (right) of a ‘normal’ rib shows a probable solution for the fixing of gunwale to rib, to plank, based on the evidence provided by examination of the rivets, their lengths and the angle of the roves. Note that the diagram, taken from Volume 1, includes the thole and its securing spike. The top edge of the gunwale is halfway down the gunwale spike.

  • A gunwale-level, rudder-support beam would need securing to the craft
  • The support ribs 24 & 25 are angled backwards as if they should function as support struts or brackets to a load at gunwale level. Support outside the gunwale could resemble the Roman system of brackets.
  • the lower shaped edge of the expanded top of the rib could provide a partial anchor for the securing system
  • The lower edges of the clinker planks on the outside of the ship might provide purchase for a securing system. All the internal ribs appear to have been shaped on their lower edges to follow the contours of the planking. This is a very challenging process from the point of view of accuracy, but it is much simpler to do on the outside of the craft.
  • the diagram illustrates the possibility of a hole between the rib and the side of the ship that could be used to pass lashings through
  • if tholes can be secured to the gunwale with spikes, could rudder support beam(s) also be secured in this way? This would be a simple method. Most other options involve fabricating relatively complex structures.

Gunwale spikes

Information relating to the gunwale can be gleaned from Volume 1. In general, it was quite difficult to record the gunwale exactly as it was the highest and most vertical part of the craft subject to the worst effects of the weather. After the war the impression had generally disappeared along with the top 3 planks and 5ft-6ins of the stern had also been destroyed by mechanised vehicle activity. Both ends of the ship were relatively narrow and subject to extra damage.

Page 398 explains that ‘beyond rib 23 and rib 4 the strake (gunwale) had been destroyed’. Our interest is at rib 24, just one frame into the damage. We are so near, but so far. There is an illustration (half scale) of a gunwale spike on page364, Fig 279; it is described on page 405 as being up to 7ins long. The spikes were used to secure the tholes to the gunwale and they were arranged in a distinct regular pattern. They appear in photographs from the 1939 dig and they were described. The series of thole spikes ran to rib 23 where, it will be seen in our first illustration at the beginning of this investigation, that the tholes ended. So near, again, to learning more about rib 24, but the ship would have been too narrow at this point for a rower and astern of this point would be reserved for the steering operation. On page 405 it is described like this: ‘The series then runs aft to rib 23, with a single gunwale spike at rib 24 (fig 312)’. The figure is on page 407 and is traced from the photograph by Phillips on the previous page mentioned at the beginning of this article. There is no discussion of this isolated gunwale spike. Why should it be there? What was its function? It opens the possibility of securing something to the gunwale in this area.


Steering 1939

The 1939 dig was done under extreme pressure in the shortest possible time. Volume 1 reports:

‘A note in Phillips diary 16 August 1939 says that this was the day when the steering system was originally identified’ and another note on 23 August described the search for the button on the outside of the ship. The excavating ended 24 August 1939 and the site was cleared the next day. Volume 1 says ‘the method of the attachment of the rudder to the ship is one of conjecture.’ The conjecture must now end and a viable system must be produced consistent with what was discovered.


Rudder Blades. The Nydam rudders

Nydam preceded the Sutton Hoo ship by 300 years, but in many ways the Oak ship resembles it quite closely. There was also a Pine ship. The rudder found with the Oak ship was badly damaged and so the rudder on the current, re-assembled original craft is a replica of that rudder. It is 3.2 m long and .55m wide.

The Pine ship’s rudder was discovered in 1993 and is well rounded; its shape is similar to a modern sailing dinghy rudder. This rudder is the oldest, original and complete rudder found in Northern Europe. The short tiller is in line with the blade and permanently fixed. This is unusual.

The rudder has some very interesting features. The cross sections are shaped foils, that is, they have a thick, rounded, forward edge and then taper off to the trailing edge. Many rudders are this shape today. It shows an understanding of what is called lamina flow. It is the way that water ‘sticks’ to the sides of the rudder and flows away from it that is important.

A very unusual feature is a strip of rounded wood attached to the inner face of the Oak ship's rudder. Its function is open to a degree of speculation. The most obvious reason could be to act as a rubbing strip to protect the blade if it should be swept inwards and pressed against the hull of the ship. For this purpose it is odd that it is fixed vertically. It could have been fixed horizontally to create less drag in the water. Perhaps it is vertical to provide protection as the blade could strike the hull in various places and the strip would cushion the blow over a range of locations?

Another function could be hydrodynamic. When water or air flows over a foil of this shape, it will have an even effect on both sides of the foil whilst it is facing directly into the water/air flow. When it is turned, the water flow becomes uneven on each side; one side has more drag than the other. If the two sides of the foil are not the same shape, a similar effect is produced. The best illustration is the aeroplane wing which is basically curved on top and flat underneath. This changes the pressure of the air that then sucks the curved surface upwards. On a rudder that has an obstruction on the inside such as that created by the vertical wooden strip, there would be a tendency for the blade to be constantly sucked outwards away from the ship by the water.

                If the rudder is being pulled away from the ship it will need to be restrained. On the Nydam rudder there is a hole above the strip to allow a piece of rope to be threaded through the rudder and led back to the craft. The keel here has two holes in it. It would be possible to pass a rope through one hole and lead it up to the deck on that side of the ship. The rope could then be used to control the vertical alignment of the rudder blade, yet still allow the full range of movement required of the blade. On pictures of Nydam ship reconstructions in the early 20thC the hole is being used to secure restraining ropes fore and aft on the starboard side. The rudder can be pulled forward to enter the water, or it can be lifted clear of the water by the rear rope. There is very little space behind the rudder on the Sutton Hoo ship to facilitate this lifting. The ability to allow the rudder to rise is essential if the ship is to be beached, if it runs aground, or if something needs clearing from being snagged on the blade. The rudder in the Nydam ship display has a restraining rope across to the hole through the keel. The securing at the gunwale was for illustration only, as was the angle forward, due to the supporting frame.

Ideally rudders need two securing points to the hull, one at gunwale level and another lower one. Ships of the Viking era did this. The Roman ships sometimes only had a single securing point at gunwale level. The Nydam Oak ship might have had the ability to steady the blade at a lower level using hydrodynamic force and an adjustable restraining rope.

                                                                    Right. Nydam Schleswig-Holstein

The rudders are balanced. This means that there is some rudder area in front of the vertical pivoting axis. This reduces strain on the rudder and on the person holding the tiller. The more vertical the rudder the more effective this becomes. The area of the rudder ahead of the axis should be in proportion to the area behind it. Nydam rudders are low aspect ratio, which means broad. Rudders such as these are usually used on slowish craft in shallow waters. The Viking craft had higher aspect ratios, meaning slender, longer rudders as are used on swifter (sailing) craft.

Illustrations of Scandinavian rudders mostly using button-type systems can be found on page 411, Volume 1. A note below the illustrations draws attention to the thickening of the gunwale at the contact point of the rudder. The shape of the bottom of the rudder blades is of interest, especially the extended rear bottom corner. This is visible in Kvalsund 1 which was dated 490AD +/- 70 years, so could be roughly contemporary with Sutton Hoo Mound 1. Shaping such as this, results from observing the behaviour of the rudder and the entrainment of air, in clear water. It is likely that the Sutton Hoo rudder blade would have resembled the Nydam oak rudder which does not have this shaping.


A Scandinavian Rudder as photographed by Joe Startin.

Here we see the rudder system on a smaller craft Gislinge constructed at Roskilde that has the component parts of a functioning system for sailing. The rib is greater in size on the starboard side than on the port side. The internal frame supports the shaped guiding block for the rudder stock on the outside of the craft. This block is riveted in place.

Could the much larger Sutton Hoo ship have had a similar guide block, but supported across two frames? It could have been secured with trenails. It wasn’t found. The second support outside the hull is a button system. There are associated lashings passing through the hull and wedges and securing points. These were not found.


The possibility is raised that these are all small components that do not show easily as impressions in the soil. The boss may have been removed? The excavation was hurried. Should we just simply recognise that this was probably the system used and attach similar blocks to the Sutton Hoo ship?


This is the background to considerations relating to the rudder. Do you have a potential solution? If you go to Investigation 12 Rudder Solution under Ship, on this website, there is even more relevant information.