Background
Within its current project on Coastal Defence, the Centre for Maritime Archaeology has undertaken investigations into structures associated with the defence of the Danish coastal regions. In 1994 the main focus of attention was the Hominde barrage in the dammed-up part of Rødby Fjord on Lolland, and in 1995 it was decided to embark on an investigation of the Kanhave Canal on the island of Samsø.
Fig. 1: Aerial photo of the Kanhave Canal. At the bottom of the picture traces are visible of the first of the excavations in 1995 (the light sand area at the west end of the canal), and at the top the second excavation of the year can be seen in process. Photo Werner Karrasch.
The Kanhave Canal is one of the most-discussed constructions in Danish archaeological circles. It consists of a dug canal which runs across Samsø at the narrowest point of the island, from Sælvig Bay in the west to Stavns Fjord in the East (fig. 1). The canal is about 500 m long, 11 m wide, and in some sections the banks are lined with a bulwark consisting of horizontal wooden planks, 2-4 planks high (fig. 2). The heavy timbers are pierced by square holes and fastened through these by long wooden pegs, and the structure is further reinforced by long pointed slanted struts driven deep into the subsoil. The canal is today largely dry and grass-covered, but is clearly visible in the topography on the spot.
Fig. 2: Hans Stiesdal's reconstruction of the Kanhave Canal, from Stiesdal 1964.
The local view of the canal's history was that it must be connected with the Swedish Wars of the 17th century, and the name Svenskekanalen (the Swedish Canal) became linked with it. It was not until Hans Stiesdal had the opportunity to dig a cross-section through the canal in 1960 that it became clear that this was in fact a structure of much earlier date (Stiesdal 1964). The first Carbon-14 dating indicated that the canal was from around 800 AD. More recently small-scale trial excavations were carried out in 1977 by Tage E. Christiansen and by Peter Birkedal in 1979. By careful analysis of the year-rings in the samples of wood extracted, Kjeld Christiansen was later able to determine that the wood used in the canal was felled in the year 726 AD. It is possible that there were later repairs, in the 740-50s. The timber has been identified as oak, beech, aspen and lime (Christiansen 1995).
1995 Excavation I
The author of this article became acquainted with this area during an excavation in April 1995 (Kanhave 1995, I), carried out in advance of cable-laying by Jysk Telefon across the western extension of the canal, about 25 m inland from the present coast.
At this end of the canal the low ramparts which flank it can be traced up to a distance of 50-60 m from the coast. The excavation area covered 20 x 2.5 m, and it could be demonstrated that there were no traces of the Kanhave Canal so close to the Sælvig Bay. At this part of the bay the coast has silted up and the deposited beach ridges have clearly extended the shore, moving it westward, since the Viking Age. During the excavation pumps were used to extract the ground-water, and a degree of collapse of the sand-walls was then inevitable. In order take measurements for this excavation, and for the purposes of future investigations, six points were established with Geodetic-Institute levels along the north side of the canal.
With no fewer than four investigations, and a clear dating, the Kanhave Canal could thus be considered well-explored, but there were still some unanswered questions. Further information about its construction could only be obtained, however, by a change in excavation-method. If one could only for a brief period hold the high ground-water level down, and force the sliding sand to stay in the profile walls, the information would be accessible. From the point-of-view of excavation techniques, the water-pumping system which draws water into an open ditch will also draw sand with it, and a collapse will take place in the course of the first hours or day after the opening of the ditch. The only solution is to suck water away from the earth walls and thus make the sand stay in place. Only a suction-pipe system can fulfil these criteria, and it was agreed to use this equipment when the next opportunity to excavate arose.
1995 Excavation II
On 14 August 1995 an 18-day research-excavation in the canal was begun. For the investigation a site immediately to the west of Nordbyvej (see photo) was chosen; the canal was thought to be well-preserved there. The aim of this research-investigation was to establish a complete documentation of the soil-profile of a cross-section through the canal. Above all it was thought desirable to find out whether there had been other phases than the known construction-period from 726 AD. Was the canal a short-lived event, or were there signs of longer-term maintenance? Another objective was to take measurements from the newly-established levels and coordinate-points on the site to determine the bottom-level of the canal, and to obtain information on the water-level during the period of use of the canal. This latter point is of particular significance for evaluation of the coast-line of Stavns Fjord in the 8th century and for further research in the fjord area itself. In addition an analysis was to be carried out of the soil-layers which were found in the profile-wall running across the course of the canal, for use in evaluation of the forms of silting (water- or wind-deposits, marine- or non-marine-deposits, etc.), and of whether the canal had actually been open and water-filled from time to time. Finally, analyses of the low ramparts on the north and south sides of the Kanhave Canal were to be undertaken. Was this material dug up and laid directly on the banks, and was it possible to discern different phases?
Some of these questions were directly answered during the excavation, some of them need further analysis before a reply can be given, but best of all were the unexpected pieces of additional information which turned up while the work was in progress.
A ditch 3.5 x 22 m was dug down to the level of the ground water, and 20 suction pipes, 4 m in length, were put down, partly in and partly outside the ditch (fig. 3). The excavation team could then without effort shovel their way down, first to the well-preserved boards on each side of the canal, and then down through all of the material in the ditch to a level under the present Danish water level.
Fig. 3: Kanhave 1995, the site of Excavation II, with the draining system. Photo ANJ.
The bulwark
During the excavation well-preserved plank-cladding was found on both sides of the canal. Although no dating of the most recently found bulwark-constructions is yet available, it is possible to state now that the bulwark must have been built when the canal was constructed. The uppermost planks on the north and the south side were placed at exactly the same level, and on each side, behind the planks, was the original beach ridge.
In the northern bulwark (fig.4) there were two longitudinal planks lying firmly in place behind sturdy angled struts. The boards were 4 m long and very stout in comparison with the earlier-known part of the Kanhave Canal's structure with boards of about 2.5 m in length. Plank no. 2 from the top is a half-section of a trunk on both the north side of the canal and the south; the other planks are radially split. In addition to the sturdy struts, the planks in this section of the canal were also fastened by wooden pegs.
In the bulwark on the south side there were four rows of horizontal planks 3.5 m long and of slightly less solid character. There had apparently been some subsidence during construction which had torn the timber out of the wall at this point.
Fig. 4: The bulwark on the north side of the Kanhave Canal. Photo ANJ.
The ramparts
At the excavation site the ramparts which are otherwise visible today along the banks on both sides of the canal are only preserved on the north side. The south side must have had a corresponding rampart, however - this is clearly apparent from the topography at other points along the canal. Over the course of time a number of levellings have been carried out near the canal because of grazing cattle, and the contours of the canal may have been partly effaced as a result.
Analysis of the north rampart has not yet been completed, but to all appearances during the excavation it seemed to consist of two building-phases. Both were structures of e.g. grass or peat turf. The earliest phase is a rampart about 2 m broad and 0.5 m in height, lying on the old soil-surface and probably originating from the time the canal was in use. Only very little or even nothing of the originally dug-out material from the canal-trench itself was deposited directly on the banks. Most of the dug-out material, about 11,000 m, must therefore have been transported away, and this is possibly the explanation for the dune-like formations in the canal area but further back from the banks. The earliest rampart is not quite 1 m from the sloping side of the canal, and there are indications that the south side also had a recessed rampart which was later levelled out. This means that there would have been room for a towpath along each side of the canal. The large quantities of earth which make up a later building-phase in the north rampart cannot come from the time of construction of the canal, but further analyses are also required here. Light may be thrown on the question of the date of the rampart's two building phases as a result of pollen analysis of soil-samples taken from turf at various points.
Fig. 5: A section of the many soil-layers in the cross-section of the canal-trench. Photo ANJ.
The canal-trench
The canal-trench is clearly dug down through an old overgrown beach-ridge with a height 1.75 m above present sea-level. When the canal was constructed in 726 the trench was dug 1.9 m down from the top surface, down to a depth of about 15 cm below the present sea-level. The canal has been filled up with various layers of a total thickness of 1.65 m. At the bottom there is a sandy brown mud (silt) with cut-off branches, shaped wood and large stones; on top of that there are nut-shells and a few disintegrating bones. The layer seems to have been built up under relatively static conditions, since there are leaves and other remains of plants which would have been transported away if there had been water flowing through the canal (fig.5). It is therefore not possible to prove that water flowed from the sea during the time of use of the canal. There is much to indicate that the canal was left 'closed' at the west end, or that it rapidly became closed by material deposits along the coast. Ships which wanted to come through may have been dragged over the beach ridge at the west end. The upper deposits in the canal-trench show signs of flowing water, which means that from time to time, later, sea-water came into the canal from the west and so ran through the canal from the bay in the west to the fjord in the east.
Water-level
The question of the water-level in the canal can now be analyzed in greater detail. It is thought that in 726 the water-level reached up to about halfway up the bulwark. The upper side of the bulwark is known, and the uppermost planks on both sides of the canal have disintegrated badly. To judge from these observations the water-level was probably just over 0.5 m over what it is at present. If the water-level in the Stavns Fjord area was 0.5 m higher, parts of the land-areas around the fjord must have been covered by sea. These considerations have great importance in relation to how the sea-approaches could be monitored, but closer study of this is required because other factors, such as the transport of materials along the coast, may also have a bearing.
Closing down the canal?
More than half of the wood brought back from this relatively small excavation does not come from the bulwark, but from deposits in the canal-trench itself. To the surprise of the excavation-team, larger quantities of wood turned up towards the bottom of the canal, increasing as the investigation went on. It was evident that the wood had been shaped, and a few pieces may have come from the bulwark. The pieces of wood consisted of boards cut to different shapes, almost like ordinary building-timber, and there was one piece which was fire-damaged. In addition there were long straight branches and trunks with pointed ends which suggest that they were meant to be used in a vertical position, even though they were lying horizontally when they were found. One particular trunk was 4 m in length and placed across the canal at its deepest point. In addition quantities of shaped branches and large stones were also found at the deepest point of the canal.
Most of the wooden material was found, however, in the upper part of the earliest mud-layer and is therefore from the period after the time of the canal's construction and use; parts of the bulwark were among this material. The character of the wood in the fill does not suggest refuse. Nor does the idea of throwing refuse into a newly- and painstakingly-dug canal carry much sense. What this means is something of a riddle, but it is difficult to interpret as anything other than a conscious filling-in of the canal. The dating of this closing of the canal is awaited with suspense.
Historical background
The Kanhave Canal is an important monument from a time when the history of Denmark relies mostly on archaeological remains. The building of such a canal, as Hans Stiesdal pointed out, is necessarily part of a naval structure, whether it was intended for monitoring the sea-routes to the west and east of Samsø, or whether it had the double function of also allowing the war-ships to lie accessible in a combined harbour- and canal-structure from the beginning of the 8th century. The canal thus also reflects part of the military organisation which existed while royal power was in its infancy in Denmark.
The archaeological material from the first millennium AD is particularly rich as regards weapons and defence works. Analyses of this material can throw light on aspects of the development of military organisation. A matter of particular interest is the investigation of when a navy came into existence in Denmark. In association with the coastal defence project in the Centre for Maritime Archaeology, known or suspected naval harbours in prehistoric and early historic times are therefore being investigated. This subject is still an almost untouched area, and it is therefore especially important to establish some basic facts from the beginning about a number of circumstances connected with those localities where one can assume with reasonable certainty that the fleet was assembled. From the written sources in the Middle Ages several naval assembly-points are known by name and can be identified geographically today. Other assembly-points have to be identified by means of geographical and topographical criteria, local place-names (skib- and snekke- sites) and on the basis of coastal defence-works; the latter are also reflected in the place- name material (barrages: stig, steg, stag, etc.). Research into all references to coastal defence works in the archives of the Institute of Maritime Archaeology show that the place-names which come from poles (-stika) were associated with coastal defence works during the whole of the first millennium AD and the early Middle Ages. On the other hand the Kanhave Canal is in fact the earliest naval structure known to date with which the place-name snekke is associated; close to the canal there was a mound, now ploughed up, which was called Snekkehøj. At the time of the sea-barrages of the 11th to 13th centuries the number of place-names with snekke increased. On closer analysis these many sites can be divided up into at least three categories of naval assembly-places: 1) local assembly points corresponding to the herred (shire), 2) regional assembly points corresponding to a len (feudal region), and 3) countrywide naval assembly-points. Stavns Fjord probably belongs to the last category. Perhaps in time we shall be able to differentiate clearly between these categories on the basis of the archaeological material. A good example is that of Selsø-Vestby, where in the Viking Age there was a landing-place which probably also functioned as a fleet assembly-area (Sørensen and Ulriksen 1995: 25).
At exactly the time of the Kanhave Canal there seem to have been a number of changes taking place in military conditions in Denmark. There are strong indications that late in the 6th century a military reform was carried out which included the introduction of uniform arming of the people. This reflects a change in war-technology from the old order of battle, in which warriors had many different categories of weapons, to individual fighting in which all who fought had a sword as the main weapon. To this was added a growing proportion of mounted warriors. Denmark was at that time strongly influenced by the powerful leaders of western Europe, the Franks. In the course of the 7th century, when the Franks were in a weak position internally, there was a conspicuous change in the development of Nordic weapons towards a more local Nordic form, and at the same time there emerged a clear differentiation of the military elite. This process reached its culmination a good generation before 700 AD.
Nothing is known about the existence of a Danish fleet at this point, and there is no evidence about the shape of the warship-types of that time, apart from the large Sutton Hoo warship in England (c. 625 AD). Coastal defence structures, on the other hand, are an extremely useful source. In addition to the Kanhave Canal, with a confirmed dating to 726 AD, several early barrages are known which can probably be linked to an old fleet. The coastal-defence structure in the Schlei is dated to a few years after 734, and that structure is probably contemporary with the extension of the north and main ramparts in Danevirke, which are dated to 737 AD - 11 years later than the Kanhave Canal. The structure in the Schlei is interpreted by Willy Kramer as protection for a fleet assembly-place (Kramer 1992). Another coastal-defence structure is known from the ideal natural harbour at Gudsø Vig in Kolding Fjord. The structure was investigated by Flemming Rieck, who also interprets it as protection for a fleet-harbour (Rieck 1991; 1992). The earliest group of AD Carbon-14 datings from Gudsø Vig has most recently been reckoned to be from the period around 690-780 AD.
The development of the fleet will certainly have to be sought through traces of the progress of ship technology, but it can also be detected in a defence system against overseas threats. A single piece of information in the fragmentary preserved written source-material from this time gives only a faint picture of what threats might have lain behind the extension of the defences of the country's southern border, however. For the first time, in those centuries, the land-based military apparatus in Frankish western Europe was operating in a naval situation. In 734 the Frankish king, Charles Martel, (d. 741), who was trying to reestablish the military power of the Franks in western Europe, attacked the Frisians with a fleet (Haywood 1991: 88). The Frisian army was conquered, their military leader was killed and comprehensive plundering took place. Even though the Frisians were Martel's target, this unpleasant naval surprise may well have been one of the causes underlying the establishment of safe havens for the incipient Danish fleet in the 8th century. The fact that a Danish fleet of sea-going vessels existed at the end of that century is known both from the numerous Viking attacks outside Denmark (first mentioned in 792) and from King Gotfred's meeting with the Carolingian Franks around the year 800, when he was accompanied by a fleet of 200 ships.
Anne Nørgård Jørgensen
Bibliography:
Christensen, K. 1995: Kanhave-Kanalen. In Stavns Fjord - et natur-og kulturhistorisk forskningsområde på Samsø. Eds. H.H. Hansen and B. Aaby, Copenhagen.
Haywood, J. 1991: Dark Age Naval Power. A Re-assessment of Frankish and Anglo-Saxon Seafaring Activity, London.
Kramer, W, 1992: Ein hölzernes Sperrwerk in der Grossen Breite der Schlei als Teil des Danewerk-Baues von 737 n. Chr. Geb. Archäologische Nachrichten aus Schleswig-Holstein, Heft 3, Schleswig 1992.
Rieck, F. 1991: Aspects of Coastal Defence in Denmark. In Aspects of Maritime Scandinavia AD 200-1200, ed. O. Crumlin-Pedersen, Roskilde.
-- 1992: Gudsø Vig - En Vikingetidig Samlingshavn. 11. Vikingetidssymposium 1992. Højbjerg.
Stiesdal, H. 1964: Kanhavekanalen, Samsø. Christiani og Nielsen, C.N. Post. no.6.
Sørensen, S.A.& Ulriksen, J.,1995: Selsø-Vestby. Vikingernes anløbsplads ved Selsø. Roskilde.