The days of the double-ended ship, A.D. 200-1200

Even if we had no evidence at all that any of the Northern nations had made much progress in the art of shipbuilding before the Roman invasion of Britain we should be obliged to assume that such was the case because of the very high level that is reached by an example of Northern shipbuilding of the third century a.d. This vessel, found in 1863 at Nydam in Schleswig—once Danish, then German, now Danish again—is not only such a fine piece of work that she must be the product of centuries of experience, but is also so different from any Roman ship that her builders must have been working on an entirely separate line of development.

Roman ships, as far as our knowledge goes, were always carvel-built, with their planks fitted edge to edge so as to make a smooth surface. The Nydam boat (Fig. 35), like all other surviving examples of Northern shipping of the first ten centuries A.D., is clinker-built, with the lower edge of each plank overlapping the upper edge of the plank beneath it. There is another very striking difference, in the shape of the hull. In Roman ships the bow and the stern were quite unlike one another; in Northern ships of this period the two ends are almost exactly the same. This double-ended hull

is mentioned by Tacitus in about a.d. 100 as a characteristic of the ships of the Suiones or Scandinavians, and the Nydam boat, now preserved at Kiel, is a perfect illustration of the type.

She is about 76 feet long over all and 11 feet wide; amidships she is just over 4 feet deep, while the extreme points of her stem and sternpost rise 10 feet above the line of her keel. As a matter of fact she has no real keel

in the modern sense, but has only a centre-line plank, thicker and wider than the others, 3 inches thick and 2 feet wide amidships, but tapering off at the ends to about 8 inches wide where it joins the stem and sternpost and at the same time increasing in depth to 6 inches. On each side of this keel-plank there are five planks, the uppermost shaped so that its top edge forms a gunwale about 6 inches thick. The planks were joined to one another and to the stem and sternpost by means of iron nails riveted on the inside over washers, or 'rooves,' exactly as is done to-day. The ribs—nineteen in number—were not nailed to the planks as is done nowadays, but were lashed to projecting ' clamps ' left on the inside of the planks in the same way as those on the patdi in the Brigg boat. The construction of this boat and of some others belonging to this period will be seen in Fig. 41.

Evidently the planking must have been put together first, and the ribs, cut from naturally curved timber, must have been added afterward. The rowlocks for fourteen oars on each side are cut from single pieces of

wood (Fig. 36) and are lashed to the gunwale in such a way that it would be possible to reverse them and to row the boat the other way round. Simple 'backing water ' for a short time was provided for by having the oars held to the rowlocks by lashings passing through the holes. The 'thwarts,' or seats for the rowers, rested on the clamps of the gunwale-plank and fitted over the ends of the ribs; they were supported by uprights resting on the ribs beneath them. The rudder (Fig. 37) was shaped like a wide-bladed oar or paddle. It is not quite clear how it was fitted, but the curiously shaped handle at the top was evidently a primitive form of tiller by which it was turned. There was no trace of mast or sail or of any means of supporting a mast; this agrees with the statement of Tacitus that the Suiones did not use sails.

There were some remarkable facts about the discovery of this boat. First, holes had been cut in her planking with the obvious purpose of sinking her. Secondly, she was heavily loaded with gear, and some of this seems to have been thrown overboard. Thirdly, parts of two other boats were found near, and one of them seems to have had something in the nature of a ram at each end. Perhaps the Nydam boat was returning from some plundering expedition, and her crew, finding themselves pursued, preferred to throw away their booty and sink their boat rather than leave them for their enemies. The coins found in and near the boat date from between A.D. 69 and 217, so her date is probably somewhere about the middle or end of the third century.

For our next examples of early ships we have to turn to Norway and to skip some six hundred years. This brings us to the days of the Vikings, who were so essentially a seafaring people that their chiefs were even buried in their own private boats. Burial in great mounds is a very ancient and a very widespread custom; the Pyramids are an exaggerated example of this. In the North burial-mounds are usually simpler structures with perhaps a few stones to mark the entrance. There are, however, many cases in Sweden and in the countries to the south and east of the Baltic where the grave consists of stones carefully arranged to represent a boat, and there are other cases, particularly in Norway, where actual boats have been used as burial-chambers and have then been covered by mounds of earth.

It is to this custom that we owe most of our knowledge of the appearance and the construction of ships of about the ninth century. Many of the Norwegian mounds have been explored, and in several cases they have yielded recognizable fragments of boats. Better still, two of the mounds were found to contain boats well enough preserved to allow of almost complete restoration.

At Gunnarshaug in the island of Karmo, near Stavanger, the remains of a boat were unearthed in 1887. It was in very bad condition, but it was possible to make out that it had been a clinker-built vessel about 16 feet wide with a keel about 65 feet long. In its arrangements for rowing it differed from the Nydam boat, since its oars had worked through round holes in a plank fitted above the two heavy strips of timber (Fig. 41) which corresponded to the gunwale of the Nydam boat. There was nothing to show that this boat had been intended for sailing, but in the three best preserved ol these Norwegian boats, those from Tune, Gokstad, and Oseberg, there is no doubt on the matter, because in them the arrangements for supporting the mast are perfectly clear. These three finds all belong to the south-eastern part of the country, and may well represent a different type from the Gunnarshaug boat.

The Tune boat, from near Sarpsborg on the Glommen river, east of Oslo Fjord, need not be discussed at any great length because everything found in it is shown better in the more perfect boats from Gokstad and Oseberg. It was found in 1867 and was clinker-built, about 45 feet long on the keel and about 14 feet wide. The two ends and the upper planks had disappeared, but the elaborate ' step' for the mast was well preserved.

The Oseberg boat (Fig. 38) was discovered in 1903 about three miles above the present mouth of a small river which runs into the sea just west of the entrance to Oslo Fjord. It had been used as the burial-chamber of a woman, no doubt the wife of some great chief. After being dug out it was carefully put together again and is now exhibited in Oslo. This vessel is 70 feet 6 inches long from stem to sternpost and 16 feet 9 inches wide amidships. Her keel is 65 feet long, and her depth amidships, from the lower side of the keel to the upper edge of the top plank, is 5 feet 3 inches. She has a real keel like that ol a modern boat, 10 inches deep and

8 inches wide at the top. On each side of this keel there are twelve planks; the tenth, counting from the keel, is much thicker than the others and is really a strip of timber cut into an L-shaped section (Fig. 41). The ribs come up as far as this tenth plank, and are bolted to it and to the next plank beneath it. They are lashed to the rest of the bottom planks by means of clamps of a rather more elaborate pattern than those of the Nydam

boat, but they are not attached to the keel in any way. The planking is fixed together and to the keel with iron nails or rivets. From end to end of each rib runs a beam, and above each beam there is a pair of 'knees, which supports the two uppermost rows of planking. The top plank has round holes for the oars, fifteen in number and about 3 feet 4 inches apart.

The mast is stepped nearly 3 feet before the middle of the ship. Its foot rested in a hole in a large block of wood which lay on the keel and was slotted out to fit over two of the ribs. This block was not fixed down in any way; it was held in position simply by the slots and by two strips of wood nailed to the ribs on each side of it. About 3 feet higher up the mast was held between the two legs of a long piece of wuod something like a clothes-peg. This was fixed to four of the beams; it was curved upward in the middle, and the beam just before the mast was also curved. The mast was thus supported from forward and from the sides, but was free to fall aft. To prevent it from doing this at the wrong moment it was held firm by a third piece of wood which fitted between its after side and a step cut across the two legs of the main support.

The rudder was something like that of the Nydam boat in shape, but its blade was longer and straighter. It was carried on the right-hand side, as was always the case in Northern ships with a single side-rudder.

This is why the right-hand side of a ship is called the starboard side. 'Starboard ' is simply a very slightly changed form of ' steer-board,' the steering-side. The rudder was held to the ship's side in two places: by a strap of plaited leather near the upper edge of the planking and by a long strip of flexible root of a fir-tree, passing through a hole in the rudder-blade, through a large rounded block of oak projecting about 16 inches from the ship's side, through the seventh plank from the keel, and finally through a heavy timber which" extended from the keel to the gunwale. In the top of the rudder-head was a slot cut at right angles to the blade so that the tiller lay across the ship instead of fore and aft as is usual nowadays. The helmsman sat between the tiller and the stern and pushed the tiller away from him to turn to port or pulled it toward him to turn to starboard.

There was no real deck, but there were floor-boards fitted between the beams. Right at the ends of the boat the floor-boards were raised a few inches to form two separate platforms. No seats for the rowers were found, but they must have rowed sitting because the oar-ports are not much more than a foot above the floor-boards. All sorts of things were found in the ship. There were sledges, a wagon, bedsteads, chests, buckets, bones of horses and oxen, and many other objects. Everything possible was decorated with wonderful carving. This is not only the case with the things in the ship, but even with the stem and sternpost of the ship herself (Fig. 39). All this carving has been a means of dating the find

very exactly, and experts have concluded that the burial took place between 835 and 850 and that the ship herself was probably a few years older.

Until the discovery of the Oseberg ship the best example of a Viking ship had been the vessel found in 1880 at Gokstad near Sandefjord, only about fourteen miles south of Oseberg. This ship is also preserved in Oslo (Fig. 40), and fortunately she was in such good condition that it was possible to transport her in two complete portions which were easily joined up again. She is believed to date from about 900, and it will be seen that there are points about her which certainly seem to suggest a later date than that of the Oseberg ship.

The two vessels are not very different in size, though the Gokstad ship belongs to a rather more seaworthy type. Her keel is 66 feet long, and she is 78 feet long over all. The explanation of the much greater 'rake,' or overhang, at the ends is that the stem and sternpost are not fixed directly to the keel, as in the Oseberg ship, but are attached to other pieces which are in turn fixed to the ends of the keel. In breadth the two ships are

the same, 16 feet 9 inches, but in depth the Gokstad ship with her 6 feet 9 inches is as much as 18 inches the larger. She has sixteen planks a side, as compared with twelve. The beams rest on the frames at the tenth plank and have knees above them for the upper planks just as in the Oseberg ship, but this tenth plank is no longer cut in a special shape; it is simply a rather thicker plank than the rest and the curve of the side is quite continuous, instead of being broken at this point. This difference is shown in Fig. 41.

Naturally, with six upper planks instead of two the side above the beams is very much higher. It is, therefore , quite natural to find that the oar-ports are cut in the third plank from the top instead of the uppermost . Even so, they are much higher above the beams than in the Oseberg ship. The oar-ports are just about the same distance apart, but there are sixteen of them on each side and they are fitted on the inside with small shutters to close them when the oars were not in use. One interesting feature about this Gokstad ship was the remains of a row of shields fixed all along the side, covering the oar-ports (Fig. 42). This may have been their position in harbour and perhaps when sailing, but obviously they would have to be moved for rowing, and they can hardly have been very secure when the ship was under sail. In a general way the two boats are very much alike. Such differences as there are can easily be explained by the natural improvement in shipbuilding and by the fact that the Oseberg ship belonged to a woman, while the Gokstad ship belonged to a man, and, in fact, to a very big and very strong man, as can be told from the bones that were found in her.

As in the case of the Oseberg ship, all sorts of things had been buried in the ship, among them being no less than twelve horses and six dogs. Many of the wooden and metal objects are decorated, though not to such an extent as in the later find. There were, though, in this ship, besides all the usual household articles, fragments of three smaller boats. They were not in a condition to allow of restoration, but it can be said that their construction was similar to that of the ship herself and that at any rate two of them had been fitted for sailing. Their rowlocks were like those of the Nydam boat, but were nailed to the gunwale. The keel of the largest was more than 25 feet long, and that of the smallest was 13\ feet. The roof seen in the drawing (Fig. 40) was not a part of the ship, but was added to cover the body of the dead

chief. There were, however , carved supports for spreading an awning of much the same shape. There is no doubt as to the seaworthiness of this Gokstad ship, because a full-sized copy of her was made and sailed across the Atlantic for the Chicago Exhibition of 1892. She is said to have made a speed as high as ten knots at times on the way.

In the north of Germany there have been found remains of boats of very much the same shape as these Scandinavian vessels, but of rather different construction . In length of keel they range from 29 feet to 50 feet. The longest, with a keel 5 feet longer than the keel-plank of the Nydam boat, seems to have been about 2 feet narrower, but the others are more normal in their proportions. The remarkable thing about them is that they all have ribs cut in steps to fit the planking and fastened firmly to it with wooden pins, or 'trenails.' This rather suggests a later date, but the whole matter is very uncertain. A much bigger vessel found at Broesen near Danzig in 1874 belonged almost certainly to the fifteenth century, if not later.

One other find must be mentioned. This is a boat dug up in 1899 near Bruges in Belgium. She belonged to a very different type, for she was absolutely flatbottomed . She had no keel, and her stem and sternpost rose at a sharp angle from the flat bottom which was as much as 6 feet wide amidships. Above it there were seven planks on each side. Her total length was 47! feet, and her greatest breadth 11\ feet. The mast which was stepped in one of the cross-timbers on the bottom was 33 feet long, and there were still traces of a woollen sail left on the yard. The rudder was almost exactly like that of the Nydam boat.

Here again the date is not very certain, but the boat belongs in all probability to somewhere near the Viking period. The flat bottom was a necessary consequence of local conditions and has always been a characteristic of the ships of the Netherlands. We shall see shortly that in the time of King Alfred English shipbuilders recognized two distinct foreign types: the Frisian and the Danish. It may well be that in this Belgian boat and in those from Norway we have small examples of these two types.

Certainly the Viking ships that have been preserved 'are not specimens of the real warship. They are vessels of fifteen or sixteen pairs of oars, whereas big warships had usually from twenty to thirty and sometimes more. The Long Serpent of a.d. 1000 had thirty-four pairs of oars, and Canute's largest ship a few years later is said to have had as many as sixty. The Norse Sagas, which are writings combining history and legend, give a certain amount of information about ships, but hardly ever stoop to dimensions. Still they do frequently mention the number of pairs of oars, and from this we can get a fairly good idea of the length of the ships. On the proportions of the Oseberg ship the length of Canute's big vessel would be about 297 feet, and on those of the Gokstad ship about 307. Her actual length is said to have been 300 feet. This seems very long, but it is difficult to see how she could have been much less. The distance between oars could not be reduced, and anything of the bireme type was unknown in the North, except for three ships built in Norway in 1206. Alfred the Great built big ships in England when he was fighting the Danes. He has often been called " the Father of the English Navy." This does not mean that he was the first king to own warships, but rather that he was the first to use his ships properly by meeting the

enemy at sea and trying to prevent them from landing. No doubt Alfred's ships were more or less similar to the normal double-ended Viking type, but they were bigger than previous English ships and had improvements of the King's own devising. The chronicles tell us that "they were full twice as long as the others; some had sixty oars and some had more; they were both swifter and steadier and also higher than the others; they were shaped neither like the Frisian nor the Danish, but so as it seemed to him they would be most efficient."

Near Botley on the Hamble river, which flows into Southampton Water, there are the remains of an ancient ship of some sort. Her figurehead, which was a lion, is said to have been removed about a hundred years ago, and a good deal of her timber was cut up in 1875. A all part of her planking is preserved in the West Gate at Winchester. Her keel is said to have been as much as 130 feet long, and if she belongs to the Saxon and Danish period at all—which is very doubtful—she must have been a ship of at least thirty pairs of oars. As far as one can tell there must have been very little change in Northern ships between goo and uoo. The

ships of Harold and of William the Conqueror seem to have been simply enlarged copies of the Gokstad ship. Unfortunately, we know them only from a rather unsatisfactory source—the Bayeux Tapestry, made at least fifty years after the Norman invasion by ladies who possibly had never seen a ship at all, and who had only the ordinary materials of needlework at their disposal . Still, the ships look as if they have been based on drawings or directions by some one who did understand what he was doing. There is an interesting difference between the English and Norman ships that would hardly have been thought of except by some one who really knew the two types. All Harold's ships, with one exception, have a break in the line of the side amidships.

Fig. 43 shows this break quite clearly just abreast of the mast with oar-ports in the top plank before and abaft it. Nothing of the sort appears in the Norman ships (Fig. 44). One of William's ships is shown unshipping her mast before being pulled up on the beach, and apparently the figureheads were also removed when the ships were out of the water. At least one ship seems to have shrouds to her mast. This is an important point, because it suggests that by then it must have

been the practice to make use of a beam wind at least. As long as the mast was given only a stay and a backstay it is probable that ships sailed only with the wind more or less behind them; as soon as they tried to use a beam wind, or one at right angles to their course, it would be necessary to give the mast more sideways support, and shrouds would be introduced to do this. It may be that the earlier Norwegian ships had shrouds, but there is no sign of them or of fastenings for them in either the Oseberg or the Gokstad ship.

There is another representation of a ship of the same general type on the seal of La Rochelle in France (Fig. 45). Nearly all seaport towns have at some time used a seal with a ship on it, and these seals have been preserved on documents. The difficulty is that the date of a document does not necessarily date the seal which it carries, still less the ship on that seal. The same seal has often been used for centuries; for instance , Southampton still uses a seal made in 1587, . and the ship given there would be a poor example of a modern man-of-war. Still, by taking the earliest date which can be derived from a document, and by comparing the style of workmanship of one seal with another, it is possible to get a fairly good idea of the date of most of these ship seals. That of La Rochelle is believed to belong to the twelfth century, so it can hardly be more than a hundred years later than the Bayeux Tapestry and is quite possibly contemporary.

It shows exactly the same double-ended one-masted ship with clinker-built hull and with high stem and sternpost. Besides this it shows a very important feature about the sail. The vertical lines on the lower part of the sail are almost certainly 'reef-points,' by which part of the sail could be tied up in a bundle at the foot when it was necessary to shorten sail. Ships of the Middle Ages had two ways of varying their sail-area. Sometimes they used reefpoints , and sometimes they added or removed a piece of canvas called the 'bonnet' at the bottom of the sail. This second method seems to have been the more usual, and representations of reef-points are quite rare before the middle of the seventeenth century. They do, however, occur quite often enough for us to be sure that reefs were in use at any rate between 1200 and 1500. After that we know little or nothing of them before 1650, or thereabouts, but from then onward they have been in universal use right up to the present day.

The next two seals, and others of the same kind and period, show the final appearance of the double-ended ship at a date when she was already being superseded' by an improved design. The two illustrated are the thirteenthcentury seals of Sandwich and Winchelsea. Together they give a number of most interesting details. Sandwich (Fig. 46) shows 'castles ' built, or rather fitted, at each end and a 'topcastle' on the mast. We can see the ship's boat carried on deck amidships and a hook for cutting an

enemy's rigging. There are two prongs sticking out from the sternpost; more often these are shown in one piece on top of the sternpost, and in that form this fitting, called the 'mike,' is quite a characteristic of thirteenth-century ships. It seems to have been used for all sorts of odd jobs. Often the backstay is fastened to it, sometimes it carries a coil of rope, and in one picture it supports rope, anchor, and spears. From its shape one would guess that it was meant originally for supporting the mast when lowered. Egyptian ships had something very similar, and modern sailing-boats have a like fitting for carrying the boom. The three lumps on the side are no doubt 'fender cleats' to protect the planking from damage against a quay. One most important fitting appears here for almost the first time. This is the bowsprit projecting over the bow. This spar will be considered in the next chapter. The Winchelsea ship (Fig. 47) has no bowsprit and no topcastle, but its 'forecastle' and 'aftercastle ' are more highly developed. As it gives the starboard side of the ship it shows the whole of the side-rudder, which only just appears in the other example. It shows also an interesting detail in the way the anchor is being raised, with two men working on a windlass abaft the mast and two others hauling directly on the cable.

It is worth considering why Northern ships should suddenly have acquired so many new fittings about 1200 after having remained very much the same for several centuries. The most likely explanation is the increase of intercourse with the Mediterranean, caused to some extent by the establishment of the Normans in Italy and Sicily as well as in the North and more definitely by the Crusades. In the Third Crusade, which began in 1188, ships from the ports on both sides of the English Channel entered the Mediterranean as a fleet for the first time in history. As builders of seaworthy ships the Northerners may have had little to fear in a comparison with the peoples of the Mediterranean, but in matters of accommodation and of labour-saving devices the latter, with an older civilization behind them, would be likely to lead the way, and it is just in this sort of thing that Northern ships seem to have developed at the end of the twelfth and the beginning of the thirteenth centuries.

A few years later the Northern ship with improvements of her own returned to establish herself in the Mediterranean, and there she received the final touches that led, in the course of the fifteenth century, to the adoption throughout Europe of one standard type, the full-rigged ship.