Ships and Shipbuilding. Until the epoch-making excavation of the Skuldelev ships (Denmark) in 1962, the royal burial ship from Gokstad on the Oslofjord (Brøgger 1951) had dominated thinking about the ships of the period, and its presumed character was attributed to all ships mentioned in the sources. This association was partly because of the superbly uncompromising craftsmanship with which the Gokstad ship achieves its evident aim of a fast, necessarily ultra-lightweight, and therefore necessarily flexible, seaworthy hull. Since it is not an extreme longship war vessel like Ladby or Skuldelev 2 or 5 (Olsen and Pedersen 1967), and since Magnus Andersen's Atlantic crossing in his facsimile Viking ran 360 km. on his best day, the Gokstad ship seemed to be a fast ocean clipper and an ideal shallow-draught landing-craft combined. Although merchant ships were known from the texts, they had little influence on the common picture of the "Viking" ship. In fact, Andersen's Viking, with its continuous, watertight deck, extra stability ensured by built-on buoyancy fender, deeper keel, and quite unhistoric rig including staysails, cannot seriously be used as evidence for the original. Andersen's account is quite frank, but highly selective quotation by enthusiasts without understanding of his original text led to unrealistic adulation of the Gokstad ship. Some modern, superficially scientific enthusiasts even assert that the Gokstad ship, whose open hull puts its rail under water if it heels more than 15º, could carry full sail of 90 sq. m. in Beaufort Storm Force 10, ignoring the potentially 12-m.-high waves with their virtual gravity effect that accompany such unusual storms. Even accurate modern replica Viking ships must be treated with caution as historical evidence, for they are usually sailed by dedicated and gifted seamen with modern backup, and are not really representative of the Viking Age. But because the ship's capabilities are so central to many questions of the period, over- or underestimates can seriously affect the value of hypotheses about it.
The Skuldelev find gave a convincing demonstration on one site of the variety of Viking ship types. Skuldelev I (length 15.8 m., breadth 4.4 m., draught 1.3 m., sail area 91.5 sq. m.) was capable of crossing the North Atlantic to Iceland or Greenland with 13.6 metric tons of cargo, but was not very fast under sail nor easily rowed. The smaller Skuldelev 3 (length 13.4 m., breadth 3.5 m., draught 0.9 m., sail area 41.2 sq. m.) was equally seaworthy, with a crew of six to seven but room for twice as many, and is probably representative of the maid-of-all-work of many estates of the period. The Skuldelev war vessels 2 (length 28 m., breadth 4.4 m., draught 0.9 m.) and 5 (length 18.3 m., breadth 2.6m., draught 0.6 m.) had much larger crews, of a minimum fifty and twenty-five respectively, with 100 or fifty needed for continuous operation, requiring larger supplies of food and water scarcely to be carried in these light-displacement hulls, making them less independent, with a smaller operational radius. They are developments of pre-Viking types of the 5th and 7th centuries, such as Nydam (Akerlund 1963) or Sutton Hoo (Bruce-Mitford 1975), which carried no sail but evidently made successful North Sea crossings under oars alone. The division in the Skuldelev craft between capable sailing vessels with auxiliary oars, and primarily oared war vessels with auxiliary sails disposes of the difficulty that the Gokstad ship is underpowered under sail and oars in most believable reconstructions, leading to a variation between them of 150 percent in sail area and displacement. The warships could easily raise or lower their mast and rig when rowing to windward; in the merchant ships, the mast was a fixture, which may suggest that it was higher, with a higher aspect ratio, more efficient rig, suiting a ship with a smaller crew and less oar power to call on. But since no rigs survive, there can be no proof.
The variety of craft at Skuldelev is important in other ways. Once the Gokstad ship ceases to be the sole model, the skepticism of some historians about the size of certain ships (e.g., King Óláfr's Long-Serpent) or certain fleets (e.g., the A.D. 851 fleet of 350 ships) seems less justified. If we enlarge it to a length of 51.8 m. with eighty rowers, or envisage 20,000 men in a fleet of such expensive masterpieces, we may well have doubts. But a longship like Skuldelev 2 would at such a length have only twice the tonnage of the Gokstad ship (Banbury 1975), with the same power per ton available from its oars, and a fleet made up of Skuldelev 3 need not have numbered more than 7,000 men, a good deal more credible than the 20,000 involved if they were all in Gokstad ships. The contemporary accounts are less open to the charge of possible exaggeration.
Existing ships make a bigger impression than deductions, however legitimate, from fragments. But the evidence from Staraja Ladoga and Hedeby of merchant ships of Scandinavian type of 24.4 m. long and one from Bergen 27.4 m. long shows that some were larger than later cogs, such as that from Bremen, which is 23.2 m. long (Crumlin-Pedersen 1983). There is thus no reason to use the nature of the vessels as evidence that Viking ships carried only small luxury cargoes, and the bulk shipment of grain and other low-value loads later produced a different ship type, the Hansa cog. The development of ships of this size in Scandinavia is evidence that bulk cargoes were carried on a scale that made such ships worthwhile before the Hanseatic period.
Whether the later, larger oared warships were a similar improvement on their predecessors is doubtful. Since the maximum speed for a given power depends on waterline length, and an oarsman produces approximately 0.5 hp. and requires a yard of rail to row over (less in the later, more crowded ships), the motive for increasing length beyond the limit of structural strength of a shallow hull is clear. The hull had to be shallow for its length to enable the oars to be applied effectively without entering the water at too steep an angle. In the Gokstad ship, this ratio was combined with a seaworthy freeboard by adding two strakes of planking above the oarports, with covers to close the oarports when the ship was under sail; without this feature, the permissible heel would have been too small for sailing in a seaway.
Estimates of the ability under sail of the Gokstad ship have been naturally as various as estimates of its sail area and displacement where the variation between upper and lower estimates, appropriate to different roles, is 150 percent. Some scholars thought Gokstad always needed a following wind, while others thought it could sail 4º off the wind at 7 knots, far surpassing a modern ice-yacht! The latter is certainly a misunderstanding of Andersen's 4 points (45º), itself hard to believe even as a heading, to which must be added the optimal 10° leeway required to produce the turbulence of water flowing under the shallow keel toward the windward side, which adds a virtual keel of water, shown in tank tests and sailing with replicas. The side rudder, not very effective for large course alterations, but easily handled because it is balanced about its axis, functions as a trim tab to control this angle of attack. Two alternative tiller holes in some rudders suggest it was used in the "half-up" position to control the ship even at the last minute before beaching, which was the normal way of loading, unloading, or spending the night.
The speeds to be expected from the waterline length can be substantially exceeded. Odin's Raven, a two-thirds replica version of the Gokstad ship, very similar to Skuldelev 3, was timed at a speed of 12.5 knots, 2√L, and under a conservative seagoing rig tacked through the wind reliably and repeatedly without using oars or making stemway. At such speeds, England was barely a night and a day from Denmark, but of course such speeds imply that the wave making as the ship races through the water is not that of a displacement hull. Whether we explain this by planing, semiplaning, the flexibility of the hull, or fine sections and ultra-light weight (all have been suggested; Binns 1980), the puzzling fact is well established. Light is thrown on the necessary sail trimming by the practice in contemporary North Norwegian femboring (five oars a side) of using a priare, or three-armed sheet in the middle of the foot, used to control the curve arid therefore the power of the sail. The interlace at the foot of the sail being held by crewmen in depictions of Viking ships on Gotland stones, such as that from Hejnum, seems to show this as a Viking Age feature, which implies skilled fine tuning to get the best speed possible as close to the wind as possible.
Shipbuilding Seasoned timber has proved difficult to work as required in modem replica construction, and it seems likely that the timber was cut in the fall and moved through the forest on the winter snow to some convenient place close to the beach, where in spring it was built into a ship that sailed that summer. No hull has shown any sign of the use of a saw, and planks were split from a trunk radially, giving perhaps six narrow oak planks, or across a diameter, giving two wide pine planks. This process is important for the interpretation of hull shape, as shrinkage along the radius of a trunk is only half that across it (McGrail 1974). When allowance for this distortion is made, early ships like Nydam can be restored to much more seaworthy forms than those presently exhibited. The relative stability of split planks was required by the constructional method of all Viking Age hulls, the lapstrake or clinker form in which the hull shell is first completed from overlapping planks, and afterward strengthened by inserting frames and crossbeams. The extra flexibility needed in the devotedly and expensively lightened warships, where the thickness of the frames is carefully varied to suit the load, was achieved by tying the planking to the frames; but in merchant ships, it was fastened rigidly with metal or wooden spikes.
Ease of beaching required a shallow keel and relatively wide midship section, and seaworthiness required a curving high prow with reserve buoyancy. The ability to keep up speed with little power input required fairly fine ends so that the ship was not stopped by seas hitting a bluff bow. In the smaller craft in which the method is most at home, the crossbeams provided seats for the oarsmen. In ships the size of the Gokstad ship and larger, crossbeams were supports for a deck of loose boards on which the oarsmen probably sat on their sea chests. In the largest ships, a double row of such beams was needed. The cog was built on thicker planking secured to a skeleton or keel and straight stems, which was built first, with a fairly flat bottom giving a much greater volume on a given length, altogether more suitable for carrying large cargoes in tidal waters where oared propulsion was less important.
Alan Binns