A personal account - Paul Constantine
Planks and Scarfs
When people look at the ship drawings their major impression is the planks as defined by the lines of rivets. I mentioned in Investigation 2 that my very first concern was to see how wide the widest plank would be for this would determine the viability of the project. The wider the wood the more difficult it would be to find a suitable source. Historically, this has been the case. In chronological order, Hjortspring had 2 planks a side; Nydam 5 planks a side and Sutton Hoo 9 planks a side. Oseberg had 12 planks and Gokstad 16 planks a side. Of course, the shape and function of the ships also determined plank width. The Viking sailing ships had a wineglass cross-section involving a pair of sharply turning corners on either side and the sharper the turn, the easier it will be to make that turn using narrower planks or strakes as they are sometimes called. The Sutton Hoo ship did not have this problem. Its smoothly rounded hull gave the best chance of using approximately even-width boards in the midship area where the wider boards would be used. The Hjortspring boat and the Nydam ship both had planks that ran virtually the whole length of the hull of the craft. The boards are at their widest in the middle of the ship and then they decrease in width as they approach the ends of the ship, meaning that they are easier to source. It is not desirable to include the very centre of a tree, called the Pith, or the layer of new wood just below the bark called the Sapwood in the plank. This means that the best planks can only really be taken from one half of the tree, instead of right across it and this is why wide planks are difficult to acquire as the tree needs to be both very long and very large.
In Volume 1 (see Investigation 1) a table of plank widths appears on page 354.
The table is incomplete especially in the middle, bottom of the ship, ribs 11 - 18 that are not included above, where the excavator’s feet walked. The measurements were all taken on the lines of the ribs which are numbered on the left-hand side of the table. The keel marked 'K' is in the middle at the top and the plank (strake) widths work outwards on either side. Plank no 1 next to the keel is known as the Garboard strake. It can be seen that the planks widths were not an exact match from one side to the other. It is possible to fill in some of the blanks by looking across to the opposite side of the hull to see what was recorded there. It is also possible when knowing the length around the outside skin of the ship from gunwale to gunwale to do some basic maths to discover some plank widths. It has to be stressed that these plank widths are based upon the measurements rivet to rivet. If the rivets are fixed one inch in from the edge of the plank, then allowances has to be made for this.
This whole project is one of experimental archaeology. When we look at the available numbers relating to planks, it is clear that the first strakes on either side of the keel, the garboard strakes, are NOT the widest planks. The first two or three planks are less wide than those that are further out, such as plank 5. This is not the usual practice today, where the wider boards would probably be close to the keel and tend to diminish slightly as they progress outwards. This may be a controversial subject and it can be discussed endlessly, but the numbers show it clearly and it is something that present-day builders should note and be guided by. It cannot be ignored or excused as a recording anomaly. The craft may not be built as we would build now, but a major part of the motivation to construct it is to try to learn why the Anglo-Saxons did what they did and this is a typical situation where it is possible to try to understand what motivated them to make their decisions.
The Hjortspring boat (Investigation 3c) had planks from the lime tree and was sewn together using lime bast which comes from the inside of the layer of bark on the tree. The bark is soaked in water for some months and then the long strands of the bast can be stripped off and twisted by hand to form a strong cord. Twisting several cords together produces larger, stronger ropes.
The 25.5m Nydam ship used oak. Ships are built using green oak. When the tree is newly cut the timber is soft and malleable. It hardens as it dries, becoming stronger and stiffer. Finding oak today that can produce wide planks of close to the length of the ship is not easy. It was a relief to see that the 27m Sutton Hoo ship had planks that were joined end-to-end to make up the length of the ship. This makes the timber easier to source. The timber had been crushed to the thickness of a piece of paper and leeched away. It showed only as a black line in the soil. The grain pattern of the inside of the plank could be seen in places when newly excavated. Examining the pattern left on the oxidised rivets it was identified as being oak. A small quantity of willow is needed to make the trenails that are dowel-like pins used to hold timbers together in various parts of the ship.
The end joints that fix one piece of wood to another are called scarfs (sometimes written scarph). The 1939 drawings recorded their positions where they could be identified, but there had been so much ‘foot traffic’ in the bottom of the boat it was not a complete record, with planks 1 (next to the keel) to 4 being particularly damaged. The gunwale areas were also damaged, so the best planks to find strong evidence of the joints were planks 5 – 8. There is a table of plank joints on page 365 In Volume 1 that has been assembled from 1939 measurements, photographs and includes joints discovered in the re-excavation 1966/67. Plank joints did not coincide with ribs. The most common distance between joints is about 18ft (5.45m), but shorter planks were used.
The question of Scarfs
To secure the ends of the planks together a scarf joint is used. The length of any scarf is usually described using a ‘rule-of-thumb’ relating to the thickness of the wood. The length of a very short scarf might be 4 times the thickness of the wood, but it would be stronger if it was 6 times the thickness as roughly indicated in the diagram below. Ten times or even twelve times the wood thickness would provide even more length for the load to pass from one plank to another.
Three rivets in a vertical line secured the middle of the scarf. They were shorter than those rivets holding one plank to another. There has been some debate about how the scarf joint was made for there are two methods and it is not clear from Volume 1 which method was used. In one place, on page 364 it says ‘the ends of the planks were planed off to a simple oblique scarf’. This seems clear. The scarf being described is one where the ends of both pieces of wood being joined are simply tapered away by planing them and then one is flipped over so that the sloping surfaces touch each other. This relies on using a plane. The very thin ends produced are not strong. Fixing the scarf with a single line of rivets would not be as good as fixing with two lines of rivets closer to the ends, to hold the wood in tightly to stop the thin ends from lifting away. This kind of joint is often used today because it can be glued, thus supporting the thin ends, which was not an Anglo-Saxon option. The description of this joint (above), published in 1975 say that the evidence for this ‘oblique’ conclusion comes from the scarf-securing rivets placed in the mid-scarf position. It is difficult to understand how this level of accurate observation could be verified. The rivets were badly corroded and mainly iron oxide rather than solid metal. They were round in cross-section and relatively small in dimension. They may not have all been inserted at absolutely 90 degrees to the plank surface. Under such circumstances it is questionable that the any line across such a rivet could be discerned with the precision needed to determine the scarf type, because there is another scarf type that would also have produced virtually the same result.
The second type of scarf is the half-lap produced by cutting away half the thickness of wood in both planks to be joined, then as before, flipping one over to sit on top of the other. The length of both scarf types can be the same. If a scarf has to be made using an axe then it is much easier to make the half-lap rather than the oblique. The wood pattern on the rivet would be the same as for the oblique. The vulnerable thin end would no longer be a problem as it is thicker and it is protected in the half-lap. There is another description in Volume 1 which is open to interpretation. It is quite complex. On page 402 when discussing the gunwale it says: ‘failed to reveal any evidence for a plank-joint of similar construction to those in the rest of the ship – i.e. a halved scarf in a vertical plane joined by three one-inch rivets mid-plank’. This means that Volume 1 does not definitively say which scarf type was used for plank joints and it is difficult to see what evidence could have been produced from the wafer-thin black line in the soil that would have defined this beyond doubt. Which scarf to us is a difficult decision.
The nature of wood
Most of the significant plank measurements are recorded in Volume 1, sufficient to guide the builders. The actual difficulty in making planks lies in cutting them out of the tree using ancient methods. To go further, it is necessary to understand something about the nature of wood, which may not be appreciated by most people today, but which was well known by Anglo-Saxons who lived their lives in a timber-technology era.
A felled tree is very heavy and difficult to move because it is full of sap, which is mainly water. If the tree is allowed to dry out it will gradually begin to split, for as the water evaporates, the wood will shrink. If it is left for several years it will gradually break into many pieces. It splits in a particular manner with most splits running from the outside of the tree into the centre. The splits resemble the spokes in a bicycle wheel. They also look like the wire dividers on a dart board. The dart board pattern is a target that probably began as the end of a felled tree. The rings that go around it are the annual rings made by the growing tree. The wire dividers are the splits that run in towards the centre. Before people had more modern equipment, they used the natural splits to break down the tree by forcing lines of wedges into them to lever the wood apart, making the splits longer until the tree broke into two halves. Each half could then be split again using wedges to make quarters and these in turn could be split to make eighths. Eventually the planks produced would be like thin slices of a cake, thicker at the outside and thinner in the middle. By chipping away at the thicker side it could be reduced until a more uniform rectangular shape is produced, but it need not be exactly regular. (There is a short video of this process on the Home page of this website under the title: Axing the Oak.) The planks are arranged to overlap as they are attached to each other, so a thin edge and a thick edge can by placed together to make up a good average thickness.
Planks made by this riving or cleaving method possess several qualities that planks which have been sawn straight through the tree do not have. An alternative method such as sawing straight across the tree is a bit like feeding it through a bacon slicer. The annual rings in such a plank are curved and this causes the plank to bend into a curve as it dries, when viewed from the end. Cleaved or riven planks remain flatter as their annual rings are very short and fairly straight, thus reducing the amount that the plank will change shape as it dries. Riven planks have much less tendency to split through the plank which would let water into the vessel. In many respects riven planks are stronger and this is the reason for building with authentic methods that produce planks of this type. For a large craft there is almost no alternative to this riving method if stable strong planks are needed.
(Extracted from Anglo-Saxon Craft Construction)
Riving a log using wedges. Thomas Finderup