Navigation – Plan du site

AccueilNuméros21Studies on WoodThe Skuldelev 3 reconstruction, R...

Studies on Wood

The Skuldelev 3 reconstruction, Roar Ege: from reconstruction to retirement

Martin Rodevad Dael et Tríona Sørensen
p. 295-301


C’est en 1982 que le Musée des Navires Vikings de Roskilde a entrepris sa première reconstruction à grande échelle d’un bateau Viking, basée sur les vestiges de l’épave Skuldelev 3. Ce navire, fouillé deux décennies plus tôt, a été identifié comme une unité de transport et de commerce côtier daté du xie siècle. La réplique baptisée Roar Ege, lancée en 1984 et retirée du service en 2016, a fourni un ensemble exhaustif de données sur la durée de vie d’un navire viking. Cet article présente une biographie de la réplique Roar Ege, du point de vue du charpentier de marine et de l’archéologue, et détaille la manière dont la coque s’est détériorée et les composants impliqués dans son déclin.

Haut de page

Texte intégral

1In 1962, the remains of five late Viking-age ships were excavated near Skuldelev on Roskilde Fjord, on the Danish island of Zealand (Crumlin-Pedersen, Olsen 2002). The excavation and analysis of the Skuldelev ships has been published in depth elsewhere but warrants a brief summary in the context of this paper (Crumlin-Pedersen, Olsen 2002).

2The five Skuldelev ships demonstrate the diversity and specialisation of late Viking-age shipbuilding. The find encompasses two warships, one large ocean-going cargo vessel and two smaller cargo/fishing vessels for coastal use (Crumlin-Pedersen, Olsen 2002). The best preserved of the five ships is Skuldelev 3: a 14 m-long cargo/transport ship built in Denmark around AD 1040 (Crumlin-Pedersen 2002, p. 239). The aft section of the ship did not survive but roughly 75% of the hull remained, making it the most complete Viking ship found in Denmark to date (fig. 1).

Fig. 1: Torso-drawing of all preserved parts of the ship-find, Skuldelev 3, assembled to their original context and shape

Fig. 1: Torso-drawing of all preserved parts of the ship-find, Skuldelev 3, assembled to their original context and shape

(after Crumlin-Pedersen, Olsen 2002, p. 229, fig. 40)

3Twenty years after the initial excavation of the Skuldelev ships, the Viking Ship Museum began the first of a series of reconstructions of archaeological ship-finds, marking the commencement of a programme of experimental archaeological research that would come to define the Museum and its academic endeavours.

1. The Roar Ege project: Skuldelev 3 as an archaeological experiment

4Under the direction of a steering committee with Ole Crumlin-Pedersen as chairman, the Roar Ege project ran from 1982 to 1984 (Crumlin-Pedersen 1986a, 1986b; Andersen et al. 1997, p. 11). As with all archaeological experiments, the Roar Ege project also had a clear set of goals, which defined the overall scope of the project:

  1. the project would provide the opportunity to conduct ­scientific investigations into Viking-age shipbuilding and seafaring;
  2. through photos and film, the project would provide a unique insight into Viking ships and how they were handled;
  3. the project would create a direct link between the original ship-find as exhibited in the Ship Hall and the full-scale reconstruction for visitors to the Museum;
  4. the project would also allow visitors to the Museum to try their hand at various craftwork processes.

5From the outset, it was the stated intention of the steering committee that the research-based elements of the project would always take precedence over the education and outreach aspects of the build (Andersen et al. 1997, p. 10). As such, the steering committee also laid out the following rules, which provided the basis for the archaeological experiment (Andersen et al. 1997, p. 11):

  • a) the build should be led by a group of specialists within the fields of Viking ship research, shipbuilding, rigging and sailing, with Søren Vadstrup as stem smith, Erik Andersen as rig master and Max Vinner as helmsman;
  • b) the ship should be built using the same materials and techniques as can be seen on the original find;
  • c) the ship should be built as an exact reconstruction of the original ship, as it would have appeared when newly built;
  • d) the entire process should be documented in text and photographs;
  • e) a programme of test-sailing must be undertaken after the launch to establish the ship’s sailing capabilities, cargo capacity etc.;
  • f) the work would be concluded with the publication of the project and its results.

6There was one further important tenet in relation to the actual building process itself: the work of building the ship would primarily be carried out by a group of young people with experience from previous ship-reconstruction projects but without formal boatbuilder educations. The philosophy behind this decision was that modern boatbuilders, even those versed in traditional methods, would be unable to leave their inherent learning and understanding of tools, techniques and materials ‘at the door’. They would inevitably, consciously or otherwise, apply a 20th-century craftsman’s mindset to the task at hand, which could have an impact on the process of rediscovering Viking-age boatbuilding techniques (Andersen et al. 1997, p. 80).

7During the course of the build, the team became proficient in the use of reconstructed Viking-age woodworking tools and techniques, developing a skill set and an understanding of materials that would go on to become the foundation for all subsequent experimental archaeological reconstruction projects at the Museum’s boatyard.

Table 1: Materials involved in the construction of Roar Ege’s hull.

Table 1: Materials involved in the construction of Roar Ege’s hull.

8The construction phase lasted for two years, during which time the entire process was documented in detail. Table 1 gives an overview of the main materials involved in building a reconstruction of Skuldelev 3’s hull.

9A total of circa 54.21 m3 of green timber was required to build Roar Ege, equating to roughly:

  • ten oaks, varying in diameter from 0.45 m to 1 m, and in length from 4.4 m to 12 m;
  • two pine logs, with diameters of 0.3 m and lengths of 10.5 m;
  • two willow logs, with diameters of 0.4 m and lengths of 4 m;
  • 60 pieces of crooked oak for knees;
  • 25 pieces of crooked oak for floor timbers.

10It lies beyond the remit of this article, but it should be stated that a final tally of materials would include a far broader range of materials and resources: tar for treating the hull, charcoal for the production of rivets and roves, the wooden elements of the rigging, the ropes themselves, the sail, oars and anchor. For the purpose of this article and the examination of Roar Ege’s decline, the focus will remain on the hull itself.

11Roar Ege was launched on August 25, 1984 and the waterborne part of the archaeological experiment could begin (Crumlin-Pedersen 1986b) (fig. 2).

Fig. 2: Launch of Roar Ege, on August 25, 1984

Fig. 2: Launch of Roar Ege, on August 25, 1984

(photography Bo Nielsen, Viking Ship Museum in Roskilde)

12A volunteer boat guild was established and given responsibility for the general maintenance of the ship, as well as for sailing the vessel. This guild was largely comprised of the same group of people who had built the ship, something that instilled a strong sense of ownership and responsibility for Roar Ege. This can be seen in the high standard of documentation the guild carried out during the years the ship was in active use, noting everything from where and how far they sailed to when repairs and maintenance were carried out, and by whom.

13As with all of the reconstructions in the Museum’s collection, there is a division of labour in terms of the maintenance and repair of vessels. Boat guilds are responsible for the general maintenance – tarring, painting, small repairs to the sail and rig and so on – while any repairs demanding a more experienced hand are carried out by the Museum’s professional boatbuilders.

2. Lands, Rot and Rust: Clinker-Building and Its Inherent Complications

14Various small running repairs were carried out during the first two decades in which Roar Ege was in use. The first major repair came in 2003, when the mast, the two uppermost strakes, stringers, and several knees and futtocks had to be replaced due to rot.

15This repair encapsulates one of the major challenges associated with maintaining clinker-built boats. Despite regular attention and care, the construction method itself, the many overlapping components of the hull and the manner in which they are fastened, creates a fertile breeding ground for rot. Additionally, with the passage of time and the wearing of a relatively flexible hull construction, components fastened with treenails tend to become looser, trapping water and organic matter in between the surfaces, again, creating perfect conditions for rot.

16The other significant problem we faced was the rate at which the iron rivets rust and expand, creating cracks in the lands. Roar Ege was fastened using rivets and roves that were hand-forged from Siemens-Martin steel, produced using the open-hearth method in Daval, France (Andersen et al. 1997, p. 44). Siemens-Martin steel is a very homogeneous material, which, it could be argued, bears very little resemblance to the heterogeneous bog iron the majority of Viking Age ship’s nails were produced from (Lyngstrøm 2008, p. 9). However, the decision to use Siemens-Martin steel was made by the steering committee on the basis of the limited archaeometallurgical analyses of Viking-age ship fastenings that were available at the time. Drawing heavily on Buchwald’s analyses of iron finds from the Danish Iron and Viking Ages, the committee decided that Viking-age shipbuilders would have sought the best possible material to use when fastening their ships and that Siemens-Martin steel, with its low carbon content, would be a good match for this hypothetical material (Buchwald 2005; Andersen et al. 1997, p. 44).

3. The biography of a ship reconstruction: Roar Ege’s “Service Book”

17The sequence of maintenance and repair that was carried out on Roar Ege since the launch in 1984 has been documented both in text and photos, and represents a unique data set in terms of the experimental archaeological reconstruction of ship-finds. Many archaeological finds are reconstructed in full-scale and used in a manner similar to that in which the original find was employed, but to have this kind of detailed documentation over a time period spanning more than three decades is quite rare, and therefore of some significance in terms of furthering our understanding of Viking-age shipbuilding and seafaring.

18The many notes, database entries, photographs and sketches that constitute this documentation have been collated into what is now affectionately known around the Museum boatyard as Roar Ege’s “Service Book”. The list below summarises the primary phases of repair that were carried out on the hull, during the period from 1984-2014:

  • 1984-1994: Minor running repairs;
  • 1995: Two boards repaired and several knees replaced due to rot;
  • 1996: Repairs to treenails;
  • 2003: Mast replaced due to rot; 7th and 8th strakes and the stringers and futtocks in both sides replaced due to presence of brown rot and soft rot; ten knees and one breasthook also replaced due to presence of white rot;
  • 2004: Treenails replaced at frame stations 0, 1F and 1A;
  • 2005: One beam knee replaced; cracked sheer strake repaired; after stem repaired after a collision at sea;
  • 2008: Repairs to the rudder;
  • 2009: Section of the port side 4th strake replaced;
  • 2012: Sections of lead sheathing added to repair cracks in the garboards;
  • 2013: New hood ends for the 4th and 6th strakes due to spread of white rot from the second step of the stem; general problems with leaks in the 2nd strake; treenails replaced in the port side middle stringer;
  • 2014: 5th and 6th strakes replaced on both sides; six new biti, one beam and 17 new knees.

19As outlined above, various forms of rot, and the cracks and subsequent leaks caused by the rusting and expansion of iron rivets, are the two main culprits for the manner in which the ship’s hull has deteriorated. Table 2 outlines the quantities of materials which have been replaced in order to keep Roar Ege seaworthy.

20The majority of the repairs were carried out in a manner adhering to the original principles of the project. Others, such as the addition of lead sheathing to stem leaks below the waterline in 2012, represent a modern compromise but also indicate that further sections of planking would have been replaced, had we adhered strictly to Viking-age boatbuilding methods. The material estimate presented above therefore represents a conservative total.

4. The retirement of Roar Ege

21The last major repair of Roar Ege was undertaken in 2014. The 5th and 6th strakes were replaced, along with several of the internal hull components. It had been hoped that this repair would prolong Roar Ege’s sailing days by another decade, but this was not to be the case. Despite the endeavours of the boatbuilders, by the spring of 2016, the hull had deteriorated to such an extent that it was impossible to make the hull watertight without replacing even more planking. Such a repair was difficult to justify within the context of the archaeological experiment – it is difficult to imagine that Viking-age boatbuilders would have invested so much time and costly material in an aging ship.

22After 32 years of active service, the decision was therefore made to retire Roar Ege (fig. 3).

Fig. 3: Roar Ege was retired permanently on land in the spring of 2016

Fig. 3: Roar Ege was retired permanently on land in the spring of 2016

(photography Werner Karrasch, Viking Ship Museum in Roskilde)

5. The Archaeological Experiment Continues on Land

23Skuldelev 3 is estimated to have been between 20 to 30 years old when it was scuttled at the barrier at Pebberenden, a lifespan that tallies neatly with the 32 years Roar Ege had in active service (Crumlin-Pedersen, Olsen 2002, p. 333).

24The ship had undergone several episodes of repair during its lifetime, where sections of the planking below the waterline were replaced (Crumlin-Pedersen, Olsen 2002, p. 230). What is immediately striking is the disparity between the repairs carried out on Skuldelev 3 and those on Roar Ege. The majority of the repairs on Skuldelev 3 are represented by short sections of planking, most likely remedying damage that occurred when taking the ship on land or when anchored up in harbour (fig. 4).

Fig. 4: The planking of Skuldelev 3 with indication of repairs marked in purple

Fig. 4: The planking of Skuldelev 3 with indication of repairs marked in purple

(after Crumlin-Pedersen, Olsen 2002, p. 230, fig. 41)

25Roar Ege on the other hand, presents an opposing picture, where the majority of the repairs, apart from the sheathing added in 2012 and the general deterioration of the lands in 2016, have been located above the waterline. Roar Ege’s hull has also been much more extensively repaired than that of Skuldelev 3.

26These issues raise many questions concerning the building and use of reconstructed ship-finds, and they are now the focus of ongoing research at the Museum.

6. Preliminary Conclusions, and Questions for Future Research

27Roar Ege’s retirement in 2016 marked the end of the lifecycle of the first full-scale, sailing reconstruction of the five Skuldelev ships. However, Roar Ege’s contribution to maritime experimental archaeological research is far from over. The collation of Roar Ege’s “Service Book” is the first step in the research programme of a new reconstruction project at the Viking Ship Museum’s boatyard – “Skuldelev 3 Revisited” – which began in 2017 with a reinterpretation of the original ship-find and the commencement of a new full-scale reconstruction.

28The data represented in Roar Ege’s hull, and in the documentation of the many repairs conducted over the years, is now the subject of further study, as we seek to extract as much information as we can concerning the materials and resources required in sailing and maintaining Viking Age ships.

29However, some preliminary conclusions can already be drawn:

30– Life span of the hull: Roar Ege was retired after 32 years and would appear to have had approximately the same life span as Skuldelev 3, however, Roar Ege’s hull displays far more wear and tear than that of Skuldelev 3. The potential reasons for this are still under investigation;

31– Life span of the sail: Roar Ege was initially equipped with a linen pilot sail which was then replaced by a wool sail 1985, handwoven by a team of weavers at Fosen Folkehøjskole in Norway, under the direction of Solfrid Aune, Yngvild Andersen and Kirsti Godal (Andersen et al. 1997, p. 213). This sail has also had minor running repairs over the years but is otherwise still in perfect condition. The sail has therefore outlived the ship. The sail is currently in storage and we plan to continue using the same sail on the new reconstruction of Skuldelev 3, which we aim to launch in 2021, allowing us to continue monitoring the life span of the sail and see this aspect of the archaeological experiment through;

32– Material use: a total of circa 54.21 m3 of timber were utilised in constructing Roar Ege’s hull. A further 13.3 m3 were required to maintain the hull over a 30-year period. These totals can now be used to explore issues concerning the control and management of raw materials for shipbuilding in the Viking Age.

33The disparity between the level of repair as evidenced in the remains of Skuldelev 3 and in the work carried out in Roar Ege will continue to be the focus of research at the Museum boatyard for some time to come. It raises a great many questions, of which the following are just a sample;

34– Would iron fastenings produced from bog iron have greater resistance to corrosion, reducing the frequency of the need for replacement of planking? It is hoped that we can build small tender for Skuldelev 3 in 2021, and that this boat will be entirely fastened with rivets and roves of bog iron, allowing us to explore this issue more fully;

35– The oak we use today to build our reconstructions is faster growing than oak in the Viking Age; does this make it more susceptible to rot? And what impact does the surface treatment of the hull with tar have on the ship’s life span?

36– Our ships lie moored in the harbour, while Skuldelev 3’s keel shows evidence of wear and tear due to being dragged on land; would the hull’s rate of deterioration been different had it spent more time out of water? Should the new Skuldelev 3 reconstruction be handled differently?

37Some of these issues can be addressed promptly while others will require more long-term experimental archaeological analy­sis, underlining once again the vital importance of the role cross-disciplinary collaboration plays in this field of research. Roar Ege was built on a foundation of the shared expertise of maritime archaeologists, craft specialists and sailors. Concluding the research aspects of the Roar Ege project will also require the skills and knowledge of each of these groups, ensuring the continuation and development of maritime experimental archaeological practice at the Viking Ship Museum.

Haut de page


Andersen E., Crumlin-Pedersen O., Vadstrup S., Vinner M.
1997 Roar Ege: Skuldelev 3 skibet som arkæologisk eksperiment, Roskilde, Vikingeskibshallen.

Buchwald V. F.
2005 Iron and Steel in ancient times. København, Det Kongelige Danske Videnskabernes Selskab, Historiske-filosofiske Skrifter 29.

Crumlin-Pedersen O.
1986a “The ‘Roar’ Project,” in O. Crumlin-Pedersen, M. Vinner (eds), Sailing into the Past. Proceedings of the International Seminar on Replicas of Ancient and Medieval Vessels, Roskilde 1984, Roskilde, Viking Ship Museum in Roskilde, p. 94-103.

Crumlin-Pedersen O.
1986b “Aspects of Viking-Age Ship-Building in the Light of the Construction and Trials of the Skuldelev Ship-Replicas Saga Siglar and Roar Ege,” Journal of Danish Archaeology 5, p. 209-228.

Crumlin-Pedersen O., Olsen O.
2002 The Skuldelev Ships I, Roskilde, Viking Ship Museum in Roskilde, Ships and Boats of the North 4.1.

Lyngstrøm H.
2008 Dansk jern. En kulturhistorisk analyse af fremstilling, fordeling og forbrug, København, Det Kongelige Nordiske Oldskriftselskab Nordiske Fortidsminder Serie 3.5.

Haut de page

Document annexe

  • Table 1 and table 2 (application/vnd.openxmlformats-officedocument.wordprocessingml.document – 16k)
Haut de page

Table des illustrations

Titre Fig. 1: Torso-drawing of all preserved parts of the ship-find, Skuldelev 3, assembled to their original context and shape
Crédits (after Crumlin-Pedersen, Olsen 2002, p. 229, fig. 40)
Fichier image/jpeg, 648k
Titre Table 1: Materials involved in the construction of Roar Ege’s hull.
Fichier image/jpeg, 158k
Titre Fig. 2: Launch of Roar Ege, on August 25, 1984
Crédits (photography Bo Nielsen, Viking Ship Museum in Roskilde)
Fichier image/jpeg, 829k
Titre Fig. 3: Roar Ege was retired permanently on land in the spring of 2016
Crédits (photography Werner Karrasch, Viking Ship Museum in Roskilde)
Fichier image/jpeg, 443k
Titre Fig. 4: The planking of Skuldelev 3 with indication of repairs marked in purple
Crédits (after Crumlin-Pedersen, Olsen 2002, p. 230, fig. 41)
Fichier image/jpeg, 976k
Haut de page

Pour citer cet article

Référence papier

Martin Rodevad Dael et Tríona Sørensen, « The Skuldelev 3 reconstruction, Roar Ege: from reconstruction to retirement »Archaeonautica, 21 | 2021, 295-301.

Référence électronique

Martin Rodevad Dael et Tríona Sørensen, « The Skuldelev 3 reconstruction, Roar Ege: from reconstruction to retirement »Archaeonautica [En ligne], 21 | 2021, mis en ligne le 01 juillet 2022, consulté le 03 décembre 2023. URL : ; DOI :

Haut de page


Martin Rodevad Dael

Viking Ship Museum in Roskilde, Roskilde, Denmark

Tríona Sørensen

Viking Ship Museum in Roskilde, Roskilde, Denmark

Haut de page

Droits d’auteur


Le texte seul est utilisable sous licence CC BY 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search