Here follows the summary of: Crumlin-Pedersen, Ole & Athena Trakadas (eds) 2003: Hjortspring. A Pre-Roman Iron-Age Warship in Context. Ships and Boats of the North 5. Roskilde, 293 pages. ISBN 8785180521. Published by the Viking Ship Museum and the National Museum of Denmark.
The subject of this monograph is a weapon-offering from the fourth century BC found in Hjortspring Mose on the island of Als in southern Jutland and first published in 1937 by its excavator, Gustav Rosenberg. Special attention is paid here to the originally ca 19 m-long plank-built boat, which forms the central element of the find.
by Flemming Rieck
Considerable portions of the Hjortspring boat and the associated weapons were removed from Hjortspring Mose during peat digging in about 1880, without the find being brought to the attention of archaeologists. It was not until 1920 that the National Museum learned of the find and visited the site. In 1921-22, conservator Gustav Rosenberg, on behalf of the museum, undertook a thorough investigation of the small, ca 45 x 50 m bog, lying 42 m above sea level and 3.4 km from the coast.
In the course of the investigation in 1921, Rosenberg excavated large amounts of weapons of iron and bone, as well as shields and other wooden equipment. Because of the very soft state of the wood, both the excavation and the transport of the finds to the museum in Copenhagen had to be carried out with the utmost care. In 1922, attention was concentrated on the boat, which had been severely damaged by the peat digging, although considerable sections were preserved for a length of about 14 m with parts of both stems, indicating a north-south orientation. The excavation was concluded in September 1922, after a thorough examination of the surrounding area of the layer with finds. A limited examination in 1987 confirmed that the bog had been thoroughly investigated, and only a few fragments of spear-shafts and of one of the frames of the boat were found and recovered for radiocarbon dating.
2.1 Rosenberg’s analysis and reconstruction of the boat
by Ole Crumlin-Pedersen
Rosenberg’s careful measurements made during his excavation of the boat provided the basis for determining all its main features, which, with a few exceptions, were made of linden. The boat had been constructed as a delicate shell around one central bottom strake with two broad strakes on each side, fastened to each other by means of sewing along the overlaps of the planks. The rail, the upper edge of the top strake, had been reinforced but there was no trace of attachments for oars. Inside the boat were found the re-mains of ten rows of cleats for fastening lashings around elegantly-shaped frames, each consisting of a thwart with carved seats for two men, and an arched hazel branch held as a bent rib by vertical supports under the thwarts and a horizontal cross beam. The whole frame system, like the other parts of the vessel, had been designed with the greatest care with the aim of combining strength with lightness. The central bottom strake extended from both ends of the boat proper, forming sturdy, upwards-turning horns with oval cross-sections. At each end of the boat, a large winged stem rested on top of the bottom plank and the ends of the side strakes were attached to this. A vertical stem-locking board of oak fastened the winged stem to the bottom strake. The upper edge of the stem extended from the end of the actual boat in a long horn, T-shaped in cross-section. Remains of a support between the horns were found but the original lengths of these could not be determined with certainty. Traces of four cleats were found on the top of the stem probably for a rope, stretched from one end of the boat to the other.
The boat was also fitted with a raised deck and about 85 narrow floor timbers about one m in length, a bailer and about 16 fragmentary paddles and a broad steering oar at each stem. The two ends of the boat were identical in shape, but the southern end is assumed to have been the stern because the raised deck and bailer were found here. The find of a lump of caulking material and an S-shaped ‘cord-tightener’ indicate how the seams and caulking of the boat were maintained.
Rosenberg’s reconstruction of the boat was based on his observations of the surviving primary material alone, but he also noted that the resulting shape is very similar to several ship representations on rock carvings in Bohuslän, Sweden. The ship-technical assessment of the vessel and the final reconstruction drawings were provided by the naval architect F. Johannessen from Oslo. He calculated the original length of the boat, with horns, to be 18.6-19.6 m, its breadth 2.04 m, and its height amidships 0.705 m. The weight of the boat was calculated to be 530 kg and its draught 0.32 m with a crew of 24 men with arms and equipment.
2.2 Initial conservation and assembly
by Inger M. Bojesen-Koefoed & Maj Stief
After its excavation, the Hjortspring boat was conserved and physically reconstructed by Rosenberg. During the excavation and transport to the museum, all the boat’s timbers had been kept moist and well-supported. The conservation was carried out at the National Museum by the alum method, in which the wood was boiled in a solution of alum (K Al (SO4)2 - 12 H2O), in this case with the addition of glycerine. An individual support in the bath was made for each piece of wood with the aid of lead bands and plates. After boiling in the solution, all the ship’s timbers were dried and then laid aside for about 5 years, after which the wood was impregnated with melted beeswax. In order for the beeswax to become sufficiently viscous to penetrate into the wood, it was heated to 70-80° C, and this also made the wood so soft that it could be shaped.
A basement room in the newly-built extension of the National Museum was selected to be the exhibition hall and the parts of the boat were transferred here in 1932 and assembled in a cradle of wood, shaped to fit the hull form reconstructed by Johannessen. The exhibition opened in the summer of 1937, but because of the high humidity in the hall, problems soon developed with re-crystallisation of the alum and surface discharges of alum, glycerine, wax, etc., making the structure of the wood disintegrate. In 1965, it was finally financially possible to dismantle the boat in order to re-conserve it with polyethylene-glycol (PEG).
2.3 New conservation and display
by Inger M. Bojesen-Koefoed, Maj Stief & John Nørlem Sørensen
For the re-conservation, all the parts of the boat once again had to be packed separately and treated with great care because of the damage to the inner structure of the wood. Because of shortage of funds, the re-conserved boat was first unpacked, cleaned, assembled and remounted in the years 1986-88, and is now displayed in an exhibition hall with air-conditioning and a new form of presentation. Various possible models were considered for the remounting: the fitting of the original parts into a newly built boat hull made of wood or fibre glass, the exhibition of a copy of the boat with the surviving parts in a different colour, or the mounting of the surviving parts on an open frame of metal tubes that indicated the lines of the boat. Because of the general principle observed by the National Museum of exhibiting originals and not replicas, the last of these solutions was selected, employing thin stainless steel tubes both as an aesthetically and technically satisfactory way of indicating the shape of the boat and to serve as a support for the re-conserved boat’s timbers.
3.1 Planning and organisation
by Knud V. Valbjørn
The project for the building of Tilia Alsie, a full-scale reconstruction of the Hjortspring boat, and its subsequent sea trials, should be seen in the light of the particular social and cultural environment of the island of Als. Here a local factory, Danfoss, which produces instruments for thermostatic control, had attracted to the area particularly well-qualified people who formed a basis for a creative environment dedicated to quality. In addition, many of these employees have an interest in local history and links with the surrounding sea. The idea for building the reconstruction arose in 1990, when the town of Nordborg encouraged the development of projects that would make it and the region more widely known. In 1991, the Hjortspring Boat Guild was formed with a view to building and testing a reconstruction of the Hjortspring boat, which had been discovered within the area. This work was to be done in close consultation with archaeologists. The idea was to encourage the study of, and promote knowledge about, the boat’s historical and technical background, to test and exhibit the reconstruction, and to ensure that the reconstruction would be housed close to where the original had been found. The membership of the guild, which fluctuated between 50 and 100 persons, with ages ranging between 40 and 75, represented a wide range of skills and personal experience. The members divided themselves into a number of working groups, according to their own interests and abilities.
During the planning phases, priority was given to the following parameters: authenticity, legitimacy, quality and documentation, while the amount of time, or time-factor, was not fixed in advance. A 1:10 model was constructed of the boat. Various data relating to the initial reconstruction of the boat was entered into a computer for processing. Regular contact was established with the staff of the National Museum in Copenhagen and Roskilde, and monthly meetings of the guild were held and information leaflets distributed to all members. The work was carried out under critical, reciprocal control and was documented regularly by photography and video. Most of the external expenses incurred by the project, approximately DKK 500,000 or 67,000 Euros, were covered by local or national donations.
3.2 Initial work and training
by Knud V. Valbjørn & Hans P. Rasmussen
To build the boat, linden trees, Tilia parvifolia, with a 16-18 m-long straight trunk without side-branches and a lower diameter of 0.9 m were needed. Such trees, however, could not be found anywhere in Denmark or Western Europe. The members of the guild were therefore given the opportunity to do their introductory training by working with local linden. The tools for this work were based on finds of tools from the Roman Iron Age in Denmark including axes, adzes and chisels, since such finds from the Pre-Roman Iron Age are exceptionally few. Later inspiration was also derived from Danish Bronze-Age finds, Celtic finds from Germany and Switzerland and a Viking-Age find from Gotland.
For manufacturing the tools, ordinary iron was employed and the cutting edges were tempered. Experiments with shaping handles for the tools led to the development of ‘the Hjortspring iron’ for hewing off shingles against the grain. Originally, light adzes were probably used for hewing but this would have called for extensive training and it was estimated that there would be no very great difference between the finished results. The work of construction was done in an unheated building near Dyvig, and from here the official launching of Tilia Alsie finally took place in June 1999.
As basis for the work, the original boat in the National Museum in Copenhagen, Rosenberg’s publication from 1937, and Johannessen’s sketches and reconstruction drawing were used, and consultations took place with maritime archaeologists in Roskilde and Copenhagen.
As preparation for building parts of the boat, various cords from linden-bast were produced and the tensile strength tested. A panel of planks that were sewn together was also submitted to trials in order to determine sheer stress. During the introductory practical woodwork, paddles, shields and a thwart were made, and the tool angle for the axes was determined. In order to illuminate the many questions connected with the actual construction, it was decided to build two full-size sections of the boat, a 1.4 m-long amidships section and a 5 m-long stem section. The work with these elements played a decisive role in the determination of the quality of the guild’s continuing efforts, since many of the problems were solved and personal skills improved in the course of the 1,360 hours that were spent on these tasks. The weapons and other equipment in the find were also examined and measured and the tool group of the guild produced copies of some of the weapons.
The historical background of the find was analysed by a group in the guild who studied, among other subjects, the topographical conditions necessary for a fleet of Hjortspring-type boats to land on the island. For exhibition at the building site, a set of Iron-Age costumes was prepared based on contemporary bog-finds. Both sections of the boat and the weapons and costumes have been loaned to exhibitions elsewhere and a popular book on the history and background of the find was published locally.
by Knud V. Valbjørn
The tall linden trees for the planks of the boat were found in a small forest in Poland; from here four trees, 130-170 years old, were brought to Denmark. Two of these trees proved to be acceptable, while a third trunk had rot at the bottom, and the top of the fourth tree broke off while being felled. All four, however, were transported to Denmark, and three of the trunks were stored under water in Dyvig, while the fourth was immediately split in the workshop.
The bottom or keel element was the longest individual timber of the boat, with a length of 15 m, a breadth of 56 cm, and a thickness of 3 cm in the centre, diminishing towards the edges. In prehistory, the timber would have had to be hollowed out from the tree trunk and expanded to acquire the shape and dimensions required. Since experiments expanding two trial pieces after they had been boiled in water failed, it was decided instead to extend the width of the strake by gluing on extra pieces of planking. The winged stems were made from local linden, the first of which had to be discarded because of rot. The lower of the side strakes were made as broad as possible, while the sheerstrakes had to be increased in size by gluing on additions. The strake should have had an edge curvature of 30 cm in order to yield the shape reconstructed by Johannessen. Instead, the boat’s rocker had to be increased so that the curvature of the plank was reduced to 12 cm. On the inner sides of the strakes, a total of 198 cleats were made with square holes for the lashings from the frames, and on each of the stem-tops there were four cleats for the hogging truss. The ten frames of the boat were constructed like those in the original boat with a newly-felled, carefully bent hazel branch, oiled and hammered into position. The frames were fitted into the finished hull-shell and lashed to the cleats on the strakes.
The horns of the boat were made of curved linden branches fastened to the keel projections with a lock and a corresponding solution was selected for the upper horns. The stem plank of oak was set in as a ‘lock’ between the bottom and stem, and a shield- or disc-shaped, external support was fitted in between the horns further out. The joints were glued with polyurethane glue instead of resin, as in the original. Checking the shape and symmetry of the boat while it was being built was important. The bottom plank was held fast to supports made to fit the selected rocker shape. The stems were supported on the outside and the hogging truss between the stems helped to maintain the position of the stems and the rocker of the bottom. To aid in the shaping of the side planks, internal templates were set up a metre apart. This was not done during the construction of the original boat, but was necessary in order to make an exact copy of the shape of the original. The plank overlaps were caulked with a woollen thread dipped in a mixture of tallow and linseed oil. This saturated thread was laid in between plank in the overlaps. The sewing was done from the middle towards the ends with freshly tarred linden cords, after which the holes were caulked with ox tallow. This gave an inconveniently low friction in the joints and therefore a high-friction resin-tallow mixture was applied externally. A few cracks in the planks were repaired with battens that were sewn on, just as in the original.
The raised deck in the stern was made of three planks, the small floor timbers were laid out as foot supports for the crew, and the first paddles were produced in a standard version. In the course of the work of construction, the building group had to answer many questions, some of which were solved in cooperation with the National Museum, others on the basis of common sense. The horns have not proved to be of significance for the strength of the boat, but the forward keel ex-tension is useful when beaching the vessel. The hogging truss from fore to aft was useful in the building phase and also assisted later in keeping the boat in shape and evening out the tension in hogging conditions. The frames are clear examples of a weight-saving design that effectively transfers the weight of the crew to the hull.
The actual building of Tilia Alsie took a total of 6,500 man-hours, 85% of the time being spent on hewing wood and 15% on bending and joining elements. The guild, however, estimates that the work could have been done in half the time with a properly trained workforce. Of the original 12 tons of wood acquired for the building process, 0.5 tons constituted the finished boat.
3.4 Documentation and calculation of boat characteristics
by Frederick M. Hocker & Niels Peter Fenger
The shape of the finished reconstructed vessel was documented twice, first in May 1999 before launching and again in October 2000 at the close of the sea trials of that year. The great flexibility of the hull meant that it could vary greatly in shape, depending on its support. The two sets of measurements, however, were the same within a margin of 10 mm. The first set of measurements was carried out with the hull resting on supports placed at 2/9 distance from the ends and with the hogging truss hanging slack, while the second set was made with the boat resting on its specially-constructed trailer.
The measurements were in part done manually for some details and as a control, partly with the recording of a large number of points with total station, and finally with the documentation of contours, cross-sections and surface details with a ‘digital arm’. In addition the process was recorded with photographs and videos. The total-station measurements could be adapted for drawing in AutoCAD, while the digital-arms Microscribe and FaroArm registered points and lines directly for drawing with associated interfaces in the programme Rhinoceros. Certain problems had to be solved in connection with the work involved in producing a graphic surface model and a sheer draught of the boat. A double measurement was also made of the flexibility of the hull under lateral tension on the top of one of the stems. The result was a strong twisting of the boat, which gradually straightened itself out again after pressure was relaxed, indicating that most of the change was elastic, and not so much plastic.
To compare Tilia with Johannessen’s reconstruction drawing, the displacement had to be calculated from sheer draughts. With an empty hull (530 kg), the draught for Tilia is 14 cm, increasing to 35 cm with a cargo of 2,000 kg. In addition, the area of the wet surface has been calculated with a view to determining the resistance to propulsion at different draughts, and coefficients and constants have been calculated for comparison. In comparison with Johannessen’s drawing, the rocker is 12 cm greater and the sheer 3 cm deeper in Tilia. With a displacement of 2.5 m3, Tilia has a 5 cm-deeper draught amidships. In relation to the calculated coefficients, which reflect Tilia’s greater sheer and slightly rounder bottom in cross-section, it should not be forgotten that the flexibility of the hull offers the possibility of many different variations in shape and hence in the coefficients, according to how the hull is supported.
The metacentric height has been calculated from Johannessen’s sheer draught. For Tilia, the centre of gravity with paddlers on board was calculated on the basis of the midship section that was constructed first. The calculations showed that the boat would be very sensitive to lateral shifting of weight and this was confirmed by measurements made in 1999. On the basis of the measurements from 2000, the actual metacentric height was calculated to be rather higher than that calculated on the basis of Johannessen’s reconstruction. The most exposed position for the Hjortspring boat was sailing in seas with a wave-length of 13 m, corresponding to the length of the waterline, where the boat alternately rides on top of a wave and over the trough between the waves. In practice, there would also be the impact of waves on the sides of the boat and other varying stresses. The strains on the hull have been calculated on the basis of a simplification of the boat as a beam structure on an elastic base but the uncertainty as to the stability of the fastenings and the contribution made by the hogging truss means that the calculations have had to be simplified and they are only intended as a guide. During the trials in 2001, measurements were made of the tension in the stretched rope and the twisting of the hull under navigation in calm waters without waves. The recorded figures showed that the rope needs to be tightened continually in order for the tension to be maintained. The deformations of the hull were registered with a laser beam from amidships and plotted on a square plate in the bow. The movements were documented on video but proved that the boat under these conditions can be considered as a stiff body.
In the analysis of the boat’s hydrodynamics, the resistance to propulsion was calculated with the following components: frictional resistance, which depends on the area of the wet surface, residuary resistance caused by the formation of waves and a boundary layer, resistance resulting from the roughness of the bottom, and finally the resistance of the steering oar and the air resistance caused by the crew and the part of the hull above water. After launching, the actual resistance was measured by towing Tilia behind a motorboat. At speeds below six knots, the recorded values corresponded with the calculations but at speeds of 7 knots, the resistance was 25% higher than the calculations.
3.5 Sea trials
by Max Vinner
In cooperation with the Hjortspring Boat Guild, staff from the Viking Ship Museum and the Centre for Maritime Archaeology in Roskilde took part in sea trials with Tilia, in order to examine by means of practical experiments and with the aid of suitable measuring equipment how the boat could be used and its potential as a warship.
In 1999, trials were held for three days with a scratch crew of untrained paddlers from Roskilde and Als with a view to getting to know the vessel better. Initially, the boat was difficult to steer but with a steering oar fastened at the inner part of each of the keel-extensions, the boat came under control. The manoeuvring of the light boat with the paddles was quick and effective with acceleration to top speed in 30 seconds and braking from this speed in 5 seconds. A 360° turn could be carried out in 74-110 seconds, depending on technique and wind, with turning circles of 20-40 m in diameter. Measurements of speed over distances of 1,160 m and 320 m with a flying start gave speeds of 3.1-5.1 knots, depending on the cadence (38-50 strokes per minute), wind, and number of paddlers (8-18 men). The boat seems very unstable when empty but is reasonably stable and seaworthy with a full crew and 600 kg ballast on board, and it glides easily up onto the sand when beaching.
For the trials in 2000, a crew of elite paddlers with competitive experience in dragon-boat racing (these boats resemble the Hjortspring boat but are lighter) was used. With this crew, including a drummer to set the stroke, a speed of 7.6 knots was measured with a cadence of 62 strokes/minute over a distance of 1148 m. Over shorter distances, speeds varying between 7.4 and 8.2 knots were measured with GPS. Paddling was done with the Hjortspring Boat Guild’s ‘average paddles’, which proved too short fore and aft. Consequently, the crew was reduced to 18 paddlers, who achieved a speed of 6.7 knots with 57 strokes/minute with the Hjortspring paddles and 7.1 knots with 65 strokes/minute with their own, broad and light dragon-boat paddles. To conclude, Tilia was paddled out into open waters in 10-12 m/sec. winds, increasing to 14-16 m/sec., with the wind direction diagonally over the bow. A cadence of 55 strokes/min. and a speed of 5.5 knots was held initially, but the speed dropped to 2 knots when the waves reached a height of 1 m. In spite of unsatisfactory trimming, the boat coped well, also during the turn for the return voyage.
In 2001, the boat was tested in two, full-day voyages in warm, calm weather and with the same crew as in 2000. On the first day, the distance covered was 42.3 km with 4 hours 45 minutes spent underway and a break of 45 minutes, corresponding to an average speed of 4.2 knots (7.7 km/h.) including the break and 4.8 knots (8.9 km/h.) without the break. The second day’s voyage was 45.4 km long and it was combined with other tests, including an ‘inverted towing trial’ where, after the paddlers stopped paddling, resistance from water and air brought the boat to a standstill from a speed of 5.5 knots in 2 minutes and 32 seconds. The effect of letting some of the paddlers rest for short periods was to refresh the crew so greatly that they estimated they could have continued for another 4 hours, easily extending their distance to about 40 nautical miles or 74 km in one day.
A comparison of Tilia’s results of 7.6 knots in a sprint and 4.7 knots at cruising speed with rowing measurements for Helge Ask, the Skuldelev 5 reconstruction, of 5.4 knots in a sprint and 3.5-4.0 knots at cruising speed under oars, shows that the Iron-Age vessel is speedier than the Viking ship when paddling/rowing. Under sail with a favourable wind, however, the Viking ship has better results – and without even tiring the crew. With a potential day’s journey of 40 nautical miles (74 km), the Hjortspring boat would in good weather have been able to travel to Als from the islands Lolland or Femern in one day or from Skanör or Rügen in three days. In bad weather the crew would have been able to cope well enough at sea or take refuge on land. The boat would therefore have represented a dangerous threat to exposed coastlines and necessitated an effective local defence force.
3.6 The paddles
by Nadia Haupt & Niels Peter Fenger
The paddles, as an important functional part of the equipment of the Hjortspring boat, have been analysed in their archaeological and ethnographical context by Haupt as well as technically by Fenger.
Four of the paddles have been identified as field maple, Acer campestre, also used for some of the oars in the Nydam find, but the paddles used in the present sea trials were made of ash and linden. The old Chinese tradition for dragon-boat races, revived as a modern competitive sport with light boats manned in a similar way as the Hjortspring boat, provided a starting point for the trials. As described in Section 3.5, the well-trained elite paddlers from the kayak and canoe club Nord ensured realistic records of the speed potential and provided experience for the functioning of the paddles compared to modern, broad paddles. The initial scepticism towards the narrow Hjortspring paddles changed to acceptance, following the 2000 trials in strong winds and high seas.
Danish finds of prehistoric paddles are generally poorly published. The group embraces paddles with broad blades, some of these heart-shaped ornamented blades, as well as narrow-bladed paddles, some of which have a similar ridge down the middle of the blade as those from Hjortspring. The broad paddles were presumably used for short trips while the narrow ones were used for longer voyages.
Some images from the Bronze Age depict paddles shaped like the blades of spears, and corresponding shapes are known from miniatures in some of the men’s graves of the period. This might indicate the rank of these men in society, parallel to what is known from twentieth-century anthropological studies of Maori society.
The technical analysis presents a theoretical model for paddle propulsion for comparison with measurements from the 1999 trials. The kinematics of a paddler in action was recorded by marking points on the paddle and along the rail and record the movements over a full stroke on video from a tender. The hydraulic efficiency has been calculated to ca 0.75. In practice, lifting the paddle and taking it forward also consumes energy, giving a mechanical efficiency of ca 0.8. The energy needed at various speeds may then be calculated. Based on calculations and general considerations the advantages and disadvantages of narrow versus broad paddles would seem to balance.
by Knud V. Valbjørn & Max Vinner
The organisation of the Hjortspring Guild was developed over the first years with monthly sessions for all members discussing the allocation of the various tasks among the working groups of interested members. This structure was a crucial element in the success of the project. A task as complicated as that of building a complex boat type of the past calls for people with a wide range of skills and professional backgrounds, as well as a structure encouraging these talents.
It is estimated that the present building team might have constructed another similar boat in ca 5,000 man-hours, compared to the 6,500 hours spent on Tilia, provided the materials needed were readily at hand. A team of ten experienced boatbuilders would probably have been able to build the original boat in two or three months under such conditions. The design and excellent finish of the boat with elements of optimal strength at a minimal weight show that highly skilled craftsmen were originally in charge. At the same time the use of iron tools would have favoured productivity over that of the bronze and flint tools of previous periods.
The sea trials with Tilia have shown the robustness of the sewn linden bast fastenings, confirmed the use of the floor boards as foot supports for the paddlers and demonstrated that the narrow paddles served their purpose well. The initial stability of the boat is very low but satisfactory for the vessel when loaded. The hogging truss was useful during the construction phase and in the vulnerable hogging conditions at sea. In all aspects the Hjortspring boat represents an advanced exploitation of the available technology of the Pre-Roman Iron Age. In evaluating the results it is necessary to check whether the reconstruction has been optimally manufactured and the boat's crew has had the necessary qualifications for providing realistic results of the sea trials. There is not full accordance between the shape of Tilia and the lines given by Johannessen in the first reconstruction plan, but the more pronounced rocker and wider beam aft in Tilia are changes that have been properly accounted for and which no doubt make Tilia more authentic than the shape shown in the early drawing by Johannessen. The 'standard paddles' of Tilia did not match the original paddles in size, species of wood or weight. The original paddles would probably have given a slightly lower sprint speed but would have eased paddling on longer voyages. The well-trained sportsmen who paddled Tilia in 2000 and 2001 are probably as close as one can come today to the elite warriors of the Iron Age. As they only had a net period of four days for the sea trials, however, they only had limited experience with the boat type. Sea trials over a period of two weeks would have strengthened the evaluation of the results, especially in estimating the potential radius of action of the boat.
Appendix: A hypothetical 'Hjortspring skinboat'
by Knud V. Valbjørn
As a contribution to the discussion of the use of leather or wood for the construction of Bronze Age boats, a technologically-based suggestion is made here for the construction of a leather-clad vessel with a framework of stripped saplings continued outboard as horns similar to those seen in rock carvings. Calculations of the strength of the hull demonstrate that such a vessel is technically feasible, and the construction with the longitudinal elements.
by Flemming Kaul
4.1-2 Introduction and the deposition of the find
The Hjortspring find, which was deposited all at one time in about 350 BC, stands out from the other great weapon-deposits in Jutland and Fyn (those from Illerup, Vimose, Thorsberg, Ejsbøl and Nydam), all of which represent repeated deposits of weapons and date to the period 200-500 AD. These later finds have all been interpreted as offerings of war-booty from defeated armies. When the Hjortspring boat was deposited in what was then a small pond on the island of Als, it was sunk and some of the weapons destroyed, while others were plunged into the bottom of the pond. Many fist-size stones were found in the find-layer, together with bones of a horse, dog, lamb and calf. Three separate, small finds of animal bones have been interpreted as traces of earlier offerings at the site.
4.3 The weapons
The weapons comprise the largest individual group of finds. There are 169 spearheads, 11 swords, remains of several chain mail coats and parts of more than 50 shields. Of the spearheads, 138 are of iron and 31 of bone or antler. The iron spearheads have been classified by Randsborg (1995) into groups: 65 from spears, 64 from ordinary lances and eight from lances of an exclusive type. Of the swords, eight were more or less complete and three fragmentary. All the swords are single-edged with lengths of between 33 and 70 cm and with the tang lying in the central axis of the blade. Four of the swords had sheathes of ash. The small number of swords found indicates that only a minority of the soldiers carried them. The swords from Hjortspring are the oldest of their type known from Scandinavia and their shapes suggest that there were as yet no standard dimensions for this type of weapon within the Baltic area. The shields in the find make up the largest collection of surviving shields from the prehistoric period in Europe. Although they vary in length and breadth with an average length of 70-75 cm and a breadth of 45 cm, they all share the same basic form with rounded corners, wooden grips, and shield-bosses. Eleven or twelve of the shields differ from the others in that they are particularly narrow in respect to their lengths. Sixty-eight shield grips were found, ten of which were reserve grips. Fifty of the shields have been re-assembled so that they could be measured, while 14 others have been identified from the excavation records. Altogether there were probably about 80 shields of ‘Celtic’ shape, a type that was widespread in Europe at this period. The last group of weapons, the chain mail coats, is the least well-preserved. Rosenberg identified an area of 10-12 m2 with traces of rust and a few surviving rust flakes from small, linked iron rings. From the extent of this area, it is estimated that there were a minimum of 10-12 chain mail coats originally deposited in the bog.
4.4 Other objects
Other objects in the find include a few fittings and pins of bronze, cords with eyelet holes, the nozzle of a bellows, wooden tools, wooden discs with handles, a turned plate and five finely turned objects: three small wooden containers, a dish and a small cylindrical container. These five objects were all of a high standard of turning craftsmanship and the earliest representatives in Northern Europe of this technique in wood. Parallels in shape to the turned wooden containers are found in late Pre-Roman ceramic ware from Bornholm and Gotland, and from contemporary ceramics from around the Hamburg area. The model for this shape was probably the ceramic pyxis from Greece, dating to the fourth century BC, which was exported to the Adriatic and Black Sea regions, from where it spread as originals and copies in ceramics and wood. The turned wooden plate has close parallels in contemporary Danish ceramics, so that all the turned wooden objects are assumed to have been produced within the Baltic area as part of the equipment of the Hjortspring army.
4.5 Armaments, weapons and weapon-offerings in the Early Pre-Roman Iron Age
Although no other finds of armaments, weapons and weapon-offerings in the Pre-Roman Iron Age in Scandinavia and the Baltic region are anywhere near as comprehensive as the Hjortspring find, small finds are known from such places as Passentin in Mecklenburg, Tidevad in Västergötland, Krogsbølle in Fyn and Værebro in the north of Zealand. These finds and other material suggest that it can only have been quite recently before the Hjortspring deposition that the sword had come into use as a long battle-knife but that spearheads of antler and iron had been in common use throughout the entire period. In a few other finds in which parts of shields have survived, it appears that the shape of those from Hjortspring was also widespread in the area.
4.6 The date of the Hjortspring find
There has long been debate regarding the date of the find, since securely-dated diagnostic objects are not present in the find. Based on comparanda, the weapons in their totality can date the find to the fourth century BC or about 300 BC. Such a date is comparable to the two 14C-dates received from additional wooden material recovered from Hjortspring in 1987: 370 and 390 BC (390-210 BC and 400-260 BC with one standard deviation). For both samples, it is likely that they predate the felling of the tree, and thus the construction of the boat, by 10 to 50 years, and this moves the date of the find towards the middle of the fourth century BC.
4.7 The origin of the invaders
The army whose equipment was sacrified in Hjortspring Mose had probably arrived in several boats like that found in the bog. The number of weapons suggests that four Hjortspring-type boats were probably involved in the attack on Als. It is assumed that the crews of these boats represented an attacking army from abroad. On the basis of the similarity of the turned wooden containers with ceramic finds from the Elbe area near Hamburg, Randsborg has argued for a combined attack by land and sea from this region, after the vessels had been dragged across the foot of the Jutland peninsula. Kaul believes that this is less likely on the basis of the trials of Tilia, which have shown that Hjortspring-type vessels were capable of covering up to 100 km per day, making cooperation with land-troops inconvenient. The pyxis-shaped wooden containers, which require particular conditions for their preservation, might have been present throughout the Baltic region and northern Germany, since similar shapes occur in ceramics not only in the Ham-burg area but also on Bornholm and Gotland.
An attacking force, consisting then of about 100 men, would probably have been a contingent of soldiers raised by a group of 3,000-5,000 people. The find can therefore be looked upon as evidence of a development towards larger political units, evidence of which can be seen in the later bog-finds and in the later Nordic kingdoms of the Viking period.
4.8 The organisation of the Hjortspring invaders
The organisation of the invading, foreign ‘army’ is reflected in the numbers of various types of weapons found at Hjortspring, taking into account that many pieces were lost as a result of earlier peat digging. It is argued that the offering consisted of equipment from slightly less than one hundred men, each having two or three spears and a shield. Eleven or twelve of these men were distinguished by bearing a sword, narrow shields and chain mail coats. Eight bayonet-like spears and three spearheads of particularly large dimensions, which were presumably symbols of military authority, would also have belonged to this group. From experience gained from the sea trials with Tilia, it is suggested that the organisation of the army was based on four boats with two units of nine men plus one man (the paddlers on each side of the boat) and two helmsmen/‘boatswains’ and a ‘captain’ for each boat. This means that there would have been an army of 72 ‘privates’, eight ‘petty officers’, four ‘captains’ and possibly a ‘general’ or ‘admiral.’ There would thus have been eight ‘boatswains’ to watch over the boats. There is no trace of where the battle took place or its progress (except that the residents of Als were victorious) or of what happened to the defeated warriors and the three other postulated boats.
4.9 Armaments and military organisation in Europe, a short survey
In this period, there are certain similarities between the pattern displayed by the Hjortspring find and armaments in the rest of Europe. Spears and lances occur in the fifth and fourth centuries BC as the most important weapons in Greece, with the single-edged slashing sword as a supplementary weapon for some warriors. In the Celtic areas north of the Alps, the sword was practically abandoned about 500 BC. A century later, the double-edged sword made its appearance and around 300 BC this became of primary importance, supplemented by the spear. The outlined division of the Hjortspring army corresponds to the lochos/pentekostyes/enomotia-division of the Greek phalanx in its early phase around 500 BC, while the division in the fourth century BC, according to Xenophon, was based on units of a different size.
4.10 Weapon-sacrifice sites in Northern and Central Europe
Deposits of large amounts of weapons and other gear associated with armies that vary in size are known from a number of sites in Northern and Central Europe, including Scandinavia. More contemporary to but just post-dating the Hjortspring find are Celtic weapon-sacrifice sites that represent depositional activity that began about 300 BC. As opposed to the Hjortspring find, Celtic sites are repeated depositions, as are the later weapon-deposits in Scandinavia. Aside from numerous finds of weapons, all the Celtic sites also contain human skeletons, often beheaded. Al-though bodies are found in bogs in Scandinavia that date to the Pre-Roman and Early Roman Iron Age, they are not found with weapons, indicating a non-military context.
The Hjortspring find and its weapons reflect an invading, non-local army of about 100 men who attacked residents of the island of Als in the fourth century BC. The Hjortspring find also displays a development towards larger political units, as reflected in other archaeological find material from later periods. In the fourth century BC, very mobile army units of fleet-infantry transporting themselves in Hjortspring-type boats had great significance in the Baltic area and south-western Scandinavia. The extant archaeological source material reveals several possibilities for the origin of the mobile, attacking army to be in the northern German or southern Scandinavian area. These regions in the southern Baltic must have taken part in a political and military development or process that points forward towards increasingly larger political units. The later, great weapon sacrifices such as Nydam, Thorsbjerg, Vimose, Ejsbøl and Illerup Ådal, demonstrate the military and political organisation that must be assumed to have extended over large territories, and it would seem that whole provinces were capable of organising armies with a more superior and stronger leadership. The Hjortspring find would thus appear as preliminary evidence of the process that more than 1,000 years later led to the unification of the individual Nordic kingdoms.
by Flemming Kaul
5.1-2 The iconographic evidence of the Bronze Age
In this chapter experience gained from the sea trials with Tilia is used to illuminate the evidence of other archaeological sources recording the ships of the Bronze Age and the Early Iron Age, including the determination of differences between ships from the periods before and after 500 BC. The profile of the Hjortspring boat, with almost identical stems at each end, differs somewhat from those of ships in Bronze-Age iconography. The oldest of these have a downwards-turned keel-extension or ‘stabiliser’ aft and an upwards-turned keel-extension fore. Around 1300 BC, the horse-head prow first appears and slightly later, a bird-figure on the backwards-bent gunwale extension also appears; the keel-extension curves markedly upwards forward, while there is still a dipping keel-skeg aft. This shape is found with few exceptions on both bronze objects and rock carvings from the final period of the Bronze Age (Period V, 700-500 BC), while steering oars are not normally represented.
5.3 The iconographic evidence of the Early Pre-Roman Iron Age
At this time, ship representations cease to appear on bronzes but they continue in some rock-carving fields with the common feature of horns slightly curved at each end, almost parallel and identical fore and aft. Other stem conclusions, closed and loop-shape, do appear, however, but the symmetry of these fore and aft and the absence of the horizontal keel-extension aft are interpreted as evidence that they belong to the period after 500 BC. On the Dalbo II carving in south-eastern Norway, is shown a vessel with lightly upturned, parallel horns and a steering oar at each end, as well as strokes that suggest oars rather than paddles. This, the only known carving to show rowing, suggests that this technique was adopted in vessels of the Hjortspring type in the Pre-Roman Iron Age, before the changes in hull shape, visible in the Nydam boat from the fourth century AD, made themselves felt.
5.4 Transitional forms: Rock-carving ships on Bornholm
The changes in ships’ shapes from asymmetrical in profile with stabilisers aft during the Bronze Age to symmetrical in profile (sometimes with steering oars) in the Early Iron Age, probably occurred around 500 BC. Generally, no transitional iconographic forms between Bronze-Age and Iron-Age ships exist except for those in Denmark, on the island of Bornholm. These representations are depicted in the typical Bronze-Age manner with a stabiliser aft, high, raised keel-extension forward, and a horse-head stem. Based on the horse’s muzzle style, these ship depictions most likely belong to Period V of the Bronze Age, even though a slightly later date is possible. In the rock-carving field at Madsebakke, an isolated ship is depicted as a typical Bronze-Age ship with a straight hull, an extremely high keel-extension forward, a horizontal stabiliser aft, and slightly S-shaped stems which end in stylised animal heads, perhaps birds. Because of the bird heads and shape of the hull, the ship from Madsebakke can be placed in Period V of the Bronze Age. However, in the main field of the Madsebakke carvings, there are 13 closely-spaced ship representations which are difficult to assign to the Bronze Age, at least on typological grounds. A distinctive feature is the line that forms the hull, which is more curved than in other rock-carving ships in Scandinavia. Some of these representations can in fact be interpreted as a transitional or hybrid form between the typical Bronze-Age and Late Bronze Age and Pre-Roman Iron Age ship depictions. The rock carvings on Bornholm thus provide an interesting sequence of ships from those typical of the Late Bronze Age to those of the Early Pre-Roman Iron Age, which have a shape like the Hjortspring boat. In this regard, the rock-carving ships on Bornholm make an important contribution to interpreting the development of ships in Scandinavia throughout the Bronze Age as well as into a considerable part of the Pre-Roman Iron Age.
5.5 Tilia Alsie: functions, capabilities, and further considerations
The trials with Tilia have given rise to a new assessment of the function and potentialities of the Hjortspring boat. The difference demonstrated above between the Bronze-Age ships with keel-skeg aft and the Early Iron-Age ships, with identical stems and with steering oars fore and aft, can be due to the fact that the Iron-Age vessels were designed to leave the shore more quickly than Bronze-Age vessels. When landing or navigating on smaller water-courses, the paddlers could turn around and paddle in the opposite direction. This change in hull shape shows a change of ship-type from a ‘civilian’ model to a shape designed to suit the needs of a warship – a view that the Hjortspring find in its totality confirms, being the oldest, large Nordic weapon-offering find yet known.
The sea trials with Tilia show that a vessel of this type in favourable circumstances would be able to undertake long-distance voyages of up to 100 km in one day, which would have given great mobility and the possibility of surprise attacks on relatively distant shores in the Baltic area. This can in itself have promoted the formation of political defence alliances in the Early Iron Age. The Bronze-Age vessels would probably have had almost the same qualities that could have been useful for voyages to important cultic festivals, e.g. from the central Nordic habitation sites to the rock-carving fields in Bohuslän, and for voyages between the Scandinavian peninsula, Denmark and the great rivers that flow out into the Baltic, even if the weather should change underway with strong winds and high waves.
5.6 Changes in warfare, society, and politics as reflected in ship iconography The change in the shape of ships coincided with a number of basic changes in society around the year 500 BC in north-western Europe, Denmark and in the other southern and central parts of Scandinavia. The agrarian structure was also changed with the establishment of the first villages; the disappearance of iconography points to a change in religion. Objects that indicated high status and cult association lost their importance and it must be assumed that a social structure and economy based on stock breeding and the exchange of bronze was replaced by a structure based on the ownership of land and control over local resources, including iron. In this process the changes in the ships can be understood as a transformation from the vessels of the Bronze Age used in the service of the cult and exchange to the Iron-Age society’s tool for armed manifestation of power with a view to gaining control of territory.
Another change in warfare is indicated by the rock carvings at Tegneby-Litseby near Tanum, where horsemen appear with spears and rectangular shields in battle scenes. In one case, the horses have been carved on top of the depiction of a ship. Around and about 500 BC the rock-carvings thus brought the earlier cultic elements, the ship and the horse, together in a new, secular, warlike context.
by Ole Crumlin-Pedersen
In the concluding chapter, the Hjortspring boat is analysed as a primary source and set in perspective by comparison with other finds of vessels from the Bronze Age and Early Iron Age.
6.1 The source value of the Hjortspring boat
Rosenberg’s reconstruction of the boat is based on a thorough analysis of all associated parts and it has been confirmed by an examination of the find in connection with its re-conservation and the preparations for building the reconstruction. The hull form as drawn by Johannessen, however, has proved itself to be too straight in the bottom. The light and supple Hjortspring boat is thus a well-documented archaeological primary source of evidence about the standards and principles employed in boatbuilding in the Baltic region around the middle of the fourth century BC, with only the conclusion of the horns at each end remaining speculative. The longitudinal profile determined in this way reappears, as described in Chapter 5.2, in the late rock carvings, and variants of this type of vessel but without the symmetry can be traced back throughout the Bronze Age in depictions on bronze objects and rocks. The principles employed in the construction must therefore also be assumed to be an expression of a long continuity, which might even extend back to the Neolithic.
6.2 British Bronze-Age finds
Bronze-Age vessels that are comparable in size and in a wealth of constructional details with the Hjortspring boat, if not in character, are known from the British Isles in the form of the North Ferriby boats, the Dover boat and ‘the Brigg raft’ (Brigg 2), as well as several finds of parts of boats and two large logboats (Table 6.1). These finds collectively illustrate an older group of boats from the period approximately 2000-1500 BC that primarily consists of sturdy elements that had been hewn into shape before being lashed together with withes of yew. There is a younger group from the period approximately 1100-300 BC, joined together by sewing with cords that went from hole to hole, in which the individual parts were to a higher degree bent into shape. With a single exception all the boat parts are of oak and in large dimensions. There has been much speculation as to the seaworthiness of these vessels, for example in connection with the Brigg 2 find. This vessel was originally interpreted as a flat-bottomed, box-shaped barge, and later reinterpreted with the bottom curved longitudinally as well as transversally, and with bent frames stuck in through the cleats on the inside of the planks. It is difficult to determine the suitability of these vessels for navigation on the open seas because of the uncertainty as to the original shape and moulded depth. In the case of all the vessels, however, they are an expression of a well-developed wood technology, exploiting large and sturdy elements made of oak.
6.3 Danish Stone-Age boats
From the Danish area, there are a rather large number of logboats from the Stone Age and a few from the Bronze Age. C. Christensen’s survey of more than 50 Stone-Age boats dated to between 4310 and 2270 BC (uncalibrated) includes 14 boats from the Palaeolithic and 19 from the Neolithic with linden and alder respectively as the dominant species of wood. The Mesolithic boats have mainly been found along the contemporary coastline of that period. They were hewn out to a thin-walled shell with a separate transom board to close off the boat aft. Sewing has been identified in connection with repairs. At the Mesolithic Tybrind Vig site, finely worked paddles with broad, ornamented blades have been found in addition to the as much as 10 m-long logboats. The Neolithic boats are generally sturdier, with thicker walls, smaller, and in one of them irregular rows of holes survive along both the upper edges. All the Stone-Age boats have rounded bottoms and have been hewn out without inner frames or fixed transom boards. The Neolithic boats have been found in inland lakes and can therefore not be taken as firm evidence for the types of boats employed by hunters of the period or by voyagers carrying flint daggers to Norway.
6.4 Nordic boat finds from the Bronze Age and Early Iron Age
There are only a few boat-finds from Scandinavia dated to the Bronze Age. These are logboats of oak that have been found in freshwater areas so they are unlikely to have been sea-going. They differ markedly both from the Nordic logboats of the Stone Age and from the Hjortspring boat with which they share no common features, since, just like the Central European logboats of the period, they have a flat bottom and internal, transverse ridges hewn out of the trunk. From the Pre-Roman Iron Age, the only maritime find, apart from the Hjortspring boat, is a loose find of a thwart from ca 220 BC from Hampnäs in Norrland, Sweden. It is rather similar in character to the thwarts of the Hjortspring boat but comes from a boat with grown frames. Otherwise, finds date to the beginning of the Early Roman Iron Age (Table 6.2) with expanded logboats of oak from Slusegård in Bornholm and Vaale and Leck in the Schleswig region. The plank-boats from Nydam and Björke from the period ca 250-400 AD are built of oak, linden, deal and spruce, assembled with iron rivets and in some sections with the aid of sewing. These finds illustrate the changes in materials and techniques within boatbuilding in the Early Iron Age from sewn linden boats, propelled by paddling, to clinker-built rowing boats, primarily built of oak or spruce.
6.5 Comparisons between the British and Nordic boat finds
Table 6.3 shows a survey of the most well-preserved of the above-mentioned vessels with indication of their calculated hull weight in relation to the number of paddlers and the length of the boat in metres (in the case of the Hjortspring boat without the horns). Although there is some uncertainty about the figures, they do show a clear tendency for the paddled boats, where the British plank-built vessels have a calculated weight per metre and per man that lies between 124 kg and 239 kg. They are thus much heavier than the Nordic boats, where the corresponding figures lie between 21 kg and 39 kg.
This difference reflects a fundamental difference in building tradition that must have been based on the differences in both the environmental and cultural conditions. The movement of the tides around the coasts of Britain, just as in the period of the sailing vessels, would probably have played a significant role as an ‘auxiliary motor’ while sailing along the coast and in the lower reaches of the rivers, taking a great deal of strain off the paddlers. The Danish waters without tides and with the need to cut the length of many voyages by dragging the boat for a short distance across land, on the other hand, would certainly have furthered a tradition for lightly built paddled boats for coastal voyages. The Nydam boats represent the next phase, where propulsion with oars presupposes a stronger hull with a greater hull weight.
6.6 Finds from Central Europe and the Mediterranean area
The Central and West European boat-finds from the Bronze Age have no features in common with the Hjortspring boat and the vessels found in the lands around the Mediterranean are also of a completely different construction and are mainly proper cargo vessels. A few features of the Hjortspring boat are paralleled, however in iconography, such as the use of a hogging truss from fore to aft in the Egyptian ships and double horns fore and aft on a terracotta model from ca 2200 BC found in Crete. The ships in wall paintings from the seventeenth and sixteenth centuries BC at Thera (Santorini) in the Cyclades in Greece have the same lightness and elegance as that reflected in the Hjortspring boat. The shape of the hull and the age of the vessel differ but the general scenery in which the ships form part of a peaceful processional sailing, possibly with a religious background, may have provided the inspiration for the cultic processions in Scandinavia in the Bronze Age.
6.7 Hide or wood?
There has been speculation as to whether the building principle behind the Hjortspring boat was ‘skin or wood.’ There is firm written evidence of the use of hide/skin-covered boats in British waters, going back to the Roman period with the tradition continuing down to the Irish curraghs of the twentieth century. In the Celtic period, they would probably have played a significant role in local traffic on the open seas until this function was taken over in the early Roman period by sturdily built Romano-Celtic wooden ships and later by lighter, clinker-built ships in the Anglo-Saxon and Nordic tradition. For some of the ‘box-shaped’ rock-carving ships in northern Norway an interpretation as skin-covered boats is likely. For the majority of the Southern Scandinavian rock-carving ships tentative interpretations as skin-boats have been proposed on the basis of calculations of the technical possibilities, as in Section 3.8 of the present book, but these interpretations do not find support in the archaeological find-material in Scandinavia. On the contrary, the slender and elegantly formed Mesolithic paddles and logboats support the idea of an early, independent development in southern Scandinavia of lightly-built wooden boats. For lack of surviving evidence, the details in the development between the Late Neolithic period and the Bronze Age cannot be followed closely but the impetus behind the shaping of the stems may very well have come from the Greek archipelago and the Balkans on more than one occasion.
6.8 The expanded logboat
The method of production of the expanded logboat has provided a suitable model of interpretation for establishing the basic concept of the clinker-building technique with hull shape and internal support as known since the time of the Nydam boat. The discussion has focused primarily on the technical possibilities for expanding oak, but this is not difficult to do with freshly-felled oak and the technique is known over large parts of the globe for various species of wood. In the case of the Hjortspring boat, it is assumed that the unpreserved central part of the bottom plank had been expanded, and similar bottom elements probably formed the starting point for the plank boats of the Bronze Age and the Neolithic period. The slender hull-shells of the Tybrind logboats would also undoubtedly have required some degree of expansion in order to acquire the necessary stability. The stretched rope between the stems in the Hjortspring boat could have been a necessary aid in the building process, before the frames were fitted, and would have been able to hold the plank-shell and the stems in their desired expanded forms. The rope from fore to aft would likely also have been exploited to even out the strains on the hull when in heavy seas.
The model sketched here for the employment in Nordic boatbuilding of light, expanded wooden constructions in the long period stretching from the Stone Age to the Iron Age fits with the few known finds but it remains mainly hypothetical as long as sea-going Nordic ships from the Neolithic period and from the Bronze Age are not represented in the find material. For the time being we must remain content with having the Hjortspring boat as a unique example of an important type of boat from the Early Iron Age with deep roots in the preceding period.