Navigation – Plan du site

AccueilNuméros21Studies on WoodTimber supply for Vasa: new disco...

Résumés

Les registres d’achat des bois pour la construction du Vasa, achevé à Stockholm en 1628, révèlent que ceux-ci ont été achetés dans le sud-est de la Suède, à Königsberg, ainsi qu’à Amsterdam (Cederlund 1966). Une analyse approfondie des cernes des bois du Vasa révèle cependant une utilisation, dans la coque, de bois de provenances très diverses. Dans le cadre du projet intitulé « Les ressources en bois de l’Europe du Nord – chronologie, origine et exploitation (TIMBER) », l’étude dendrologique du Vasa est directement mise en rapport avec les registres d’achat des bois, de telle sorte que l’on peut détailler les volumes de bois provenant des divers lieux d’approvisionnement.

Haut de page

Texte intégral

1Purchase records tell us that timber for building the Vasa, completed in Stockholm in 1628, was bought from Småland (southeastern Sweden), from Königsberg and from Amsterdam (Cederlund 1966). In terms of the dendroarchaeological evidence for a northern European timber trade in the early 17th century, each of these timber sources presents different stories. It is perhaps not surprising in a country dominated by extensive forest that timber from the wider hinterland of Stockholm is exploited. Timber was purchased from south of Stockholm, from estates between Kalmar and Norrköping, and from estates around Lake Mälaren, inland from Stockholm (Hocker 2011). In light of this, why would timber also be imported from abroad?

2Let us look at the timber from Königsberg: we know that extensive export of timber, especially oak, took place from the towns along the southern and eastern Baltic coastline from the mid-14th to mid-17th centuries. The wood was most often converted into products ranging from clapboards, firewood, wainscots, boards and planks. Gdańsk dominated as a timber export hub, but other towns also supplied timber to be loaded onto ships to bring the product to the West. We see this trade in the dendrochronological record, particularly oak panels that were produced as an essential artists’ material for ecclesiastical sculpture and for fine art (e.g. Hillam, Tyers 1995; Wazny 2002; Daly, Streeton 2017). By the 17th century, we also see this so-called Baltic oak used for shipbuilding. The forests that were exploited seem to have been from a very extensive region. While several oak tree-ring datasets have been built for around the Vistula River, with its port of Gdańsk, and for historic sites around the Gulf of Gdańsk, that allow us to identify Baltic oak in Western Europe, we still do not know precisely where a large amount of this Baltic timber originally grew. Through the dendro­chronological analysis of oak panels that are the supports for paintings by many of the artists working during the time of the Dutch and Flemish Golden Age, groups of oak panels whose tree rings match well together have been identified. It was in the 1980s, after the construction of chronologies for northern Poland, that the Baltic source of this wood was identified (Baillie et al. 1985; Eckstein et al. 1986). The groups from the paintings have been dubbed Baltic 1, Baltic 2 and Baltic 3 and these can be separated probably because they are from different locations in the wider Baltic region (Hillam, Tyers 1995). As we do not have extensive oak tree-ring datasets for the vast region east of the Baltic Sea coast, we still cannot say where the forests were that supplied the art market and indeed the shipbuilding market in the 17th century. With the uniquely detailed purchase records for Vasa, stating that rough-sawn oak planks were purchased in Königsberg and Riga (Hocker 2011), we might examine, through dendrochronological analysis, for which parts of the ship these Baltic planks were used, and to which Baltic group they belong.

3Furthermore, according to the records, oak planks were purchased in Amsterdam. By the time of the building of Vasa, Amsterdam was importing all its timber. For example, Batavia, built in Amsterdam by the Dutch East India Company, and completed in 1628 (precisely the same year as Vasa) was made from timber from at least three sources: the planks are from the southern or eastern Baltic and from around Lübeck, and the frames are from Lower Saxony (Daly et al. forthcoming). If timber was purchased in Amsterdam for Vasa, where did this timber actually grow?

1. Material analysis

4Evidently, in spite of detailed historical records of timber purchases, there are many questions. Therefore, an analysis of the actual ship timbers of Vasa was initiated, with a view to identifying the origin of each timber through dendrochronology. Vasa’s hull is built of oak, while pine is used for some of the decks. The chief reason for performing extensive tree-ring analy­sis of the timbers used to build Vasa was to examine details of the specific use of timbers in the hull, from these different sources. The analysis is a major part of the project entitled “Northern Europe’s timber resource – chronology, origin and exploitation (TIMBER)”, in which scientific techniques for determining the provenance of timber (dendrochronology, stable isotopes and aDNA) are combined with archaeological and historical study. The opportunity to study the timber for Vasa allows us to directly compare the material evidence with the ­historical record of timber purchases, thus adding detail to the history of the logistics for a major shipbuilding endeavour in the early 17th century. The tree-ring analysis would also provide us with details of the volumes of timber that were supplied from the different sources and where the different timbers were used in the ship. Also, the differences in timber conversion can be correlated with the identification of source, giving us insight into timber availability from different regions. Finally, when it is possible to identify timbers from the same parent tree, some insight into the operational chain of the shipbuilders is possible.

2. Attaining measurements

5There are many ways to gain access to a cross section of historic timbers, to enable an uninterrupted series of measurements through the annual growth rings of timber, from the innermost to the outermost preserved rings. For Vasa we wished to make a minimally invasive analysis. I adopted the technique that is employed when analysing panel paintings, where the end grain of the panel is gently cleaned and the measurements are gained directly, using a magnifier, or photographed for later measurement. Particularly at the gunports on the upper and lower gundecks the cross section of the tree could be accessed for the ceilings and the outer hull planks (fig. 1). Also, by removing the oak sill, the cross sections of the sill itself and the frames, all oak, could be cleaned, photographed and later mea­sured on computer screen. At the same time that I have been working through the Vasa timbers (chiefly during 2017), a team from the Museum has also been onboard replacing all the iron bolts that secure the wreck. Because of this, several deck planks had been lifted to enable access through the decks for the bolt replacement team and their equipment. Thus, cross sections of a number of these (oaks on the upper and lower gundecks and pine deck planks on the orlop deck) could also be accessed for tree-ring analysis. Various other locations onboard Vasa allowed access to timber cross sections, but as the work progressed it became clear that this sampling strategy meant that timber in the ship’s uppers were analysed, while the lower decks were underrepresented. So, 42 timbers, chiefly in the lowest hull, were sampled by coring, using a dry-wood borer from the Rinntech company. Of these, 39 cores are analysed.

Fig. 1: One of the gunports on Vasa

Fig. 1: One of the gunports on Vasa

Access to the end grain of hull planks and ceiling planks is possible. Furthermore, when the oak sill is removed (it can be seen lying on the deck) the cross sections of the sill and the frames can also be accessed.

(photography A. Daly)

3. Results

6At the time of writing, 229 timbers (220 oak and nine pine) have been analysed, and of these, 203 timbers (195 oak and eight pine) have been dated. On the basis of the correlation between each timber’s tree-ring curve with each other (at their dated positions), four distinct oak groups can be identified (fig. 2). These have been designated Group 1 (blue), Group 2 (green), Group 3 (orange) and Group 4 (red). The pines form their own distinct group (lime green).

Fig. 2: The tree-ring curves from each timber analysed from Vasa are compared with each other (correlation)

Fig. 2: The tree-ring curves from each timber analysed from Vasa are compared with each other (correlation)

The grey shades indicate the correlation values, but in order to simplify the illustration these values are not presented. Four main groups have been identified.

7The tree-ring curves in each group are averaged, so that one curve represents each group. The provenance of each group is examined by mapping the correlation between each group average and a large network of master and site chronologies for oak for Northern Europe (fig. 3).

Fig. 3: The four maps show the correlation between the averages for the four Vasa groups and a range of oak master and site chronologies for Northern Europe

Fig. 3: The four maps show the correlation between the averages for the four Vasa groups and a range of oak master and site chronologies for Northern Europe

A – Group 1 (blue), B – Group 2 (green), C – Group 3 (orange) and D – Group 4 (red). The river data is from Lehner & Grill (2013, www.hydrosheds.org, accessed 3rd March 2020).

3.1. Provenance Groups 1 and 2

8The Group 1 average is 220 years in length, and covers the period AD 1404‑1623. It achieves the highest correlation (t-value) with a wide spread of chronologies but none of the t-values are greater than just under 6.00 (fig. 3A). It is highly possible that this group represents timber purchased in eastern Sweden, as mentioned in the written sources.

9The average for Group 2 is 296 years long and covers the period AD 1330‑1625. It is made from 47 timbers in Group 2 that form a core of highly correlated tree-ring curves in the group. This group is dating broadly with tree-ring datasets but it does not achieve correlations high enough to allow a firm identification of the timber source.

10The averages for Groups 1 and 2 match well together (t = 5.15). It is probable that Group 2 also represents trees that grew in eastern Sweden. It has proven difficult to confirm this however, because there are no extensive chronologies for oak for eastern Sweden from the period. Therefore, in the TIMBER project we are working on ways to solve this. One way is to build chronologies from the region if oaks from historical buildings can be located. We are looking to see if other scientific techniques can help, particularly stable isotopes that might indicate provenance. This technique has been applied, for example, in studies at Hedeby/Schleswig in northern Germany (e.g. Grupe et al. 2011) and on timber from ships (e.g. Rich et al. 2016; Hajj et al. 2017).

11In the meantime, as luck would have it, in 2017, remains of a large ship were excavated at Skeppsholmen in Stockholm, by Jim Hansson (see Hansson in this volume). Eleven oak and two pine samples were dendrochronologically dated by the author demonstrating that oaks for the ship were felled in spring 1613 and winter 1613‑14 (Daly 2017) and pine sheathing was added after AD 1623 (Daly 2018). This allowed the identification of the ship as Scepter, built in 1615 for Sweden’s King Gustav II Adolf. The ship ran aground in 1621, so the pine sheathing might have been a necessary addition after that (see Hansson in this volume). The tree-ring curves from the oak timbers from Scepter cross-date with many of the timbers from Vasa, particularly with Vasa Groups 1 and 2 (see table 1). This indicates one source for the timbers for building Scepter and some parts of Vasa. The written sources tell us that Scepter was built at Biskops Arnö in Lake Mälaren (see Hansson in this volume), and we might assume that the timber for Scepter was harvested around that lake. Now we know that some timber for Vasa was purchased from estates around Lake Mälaren, the same area in which Scepter was built. The tree-ring data we have extracted from the Scepter show us that Vasa Groups 1 or 2 indeed represent the timbers for Vasa that were purchased in eastern Sweden.

12Indeed these analyses also link us to another ship, Batavia, built in 1628. One single timber from Batavia (out of 53 dated timbers) matches best with the Vasa and Scepter tree-ring datasets (Daly et al. forthcoming).

Table 1: Skeppsholmen, Stockholm (Scepter)

Filenames

– 

– 

Scepter Ship all
Z207M004

– 

start

dates

AD 1413

– 

dates

end

AD 1613

Site chronologies

Z0923M01

AD 1404

AD 1623

6.04

Vasa group 1 63 timbers (Daly, forthcoming)

Z0923M02

AD 1330

AD 1625

6.01

Vasa group 2 47 timbers (Daly, forthcoming)

Z109M001

AD 1437

AD 1597

5.83

Barcode ship 1 BC01 Oslo 4 timbers (Daly 2013)

6094M002

AD 1450

AD 1660

5.66

Funder kirke later material 4 timbers (Daly 2002)

Single trees

Z092101b

AD 1424

AD 1607

7.06

Vasa ship bitt midships aft upper canon deck (Daly, forthcoming)

Z092114a

AD 1357

AD 1612

7.06

Vasa frame at GP9 D37 lower gundeck sb (Daly, forthcoming)

Z092118a

AD 1452

AD 1568

6.95

Vasa outer hull plank GP20 D16 (Daly, forthcoming)

Z0921179

AD 1470

AD 1593

6.82

Vasa outer hull plank GP20 D15 (Daly, forthcoming)

Z092150a

AD 1436

AD 1608

6.81

Vasa D51 ceiling right side GP4 ÖB ps (Daly, forthcoming)

BAT6047a

AD 1426

AD 1586

6.56

Batavia Western Australia 6047 frame (Daly et al, forthcoming)

Z092236a

AD 1502

AD 1625

6.41

Vasa D137 frame GP16 NB ps (Daly, forthcoming)

Z092148a

AD 1509

AD 1622

6.36

Vasa D49 frame GP4 ÖB ps (Daly, forthcoming)

Z092120a

AD 1477

AD 1607

6.12

Vasa D18 ceiling plank GP12 top left (Daly, forthcoming)

Z092111a

AD 1440

AD 1593

5.94

Vasa outer plank GP24 D13 upper gundeck (Daly, forthcoming)

Z092198a

AD 1454

AD 1588

5.70

Vasa D99 sill GP1 ÖB sb (Daly, forthcoming)

Z092241a

AD 1433

AD 1614

5.69

Vasa D142 frame GP20 NB ps (Daly, forthcoming)

Correlation between the oak average curve for Scepter and diverse site chronologies and individual trees.

3.2. Provenance Group 3

13The average for Group 3 consists of 41 timber elements, representing 31 trees (some trees supplied several timber components in the ship, as the analysis has indicated). The average is 291 years long, covering the period AD 1328‑1618. The map of the correlations between the Group 3 average and master and site chronologies is shown in fig 3C. Particularly high correlations (t-values) appear with tree-ring datasets for the region around Gothenburg in western Sweden. Thinking back to the written record for timber purchases for Vasa, mentioned above, there is no mention that timber was purchased from this region, nor from the adjacent territories of Halland or Bohuslän, which at the time belonged to Denmark or Denmark/Norway. It is highly probable that the timber that belongs to Group 3 are the timbers that were bought in Amsterdam. The main oak chrono­logy that allows us to identify this timber source is one built by Bråthen (1982). This covers the period AD 1225‑1720, and is based on archaeological timbers from the hinterland of Gothenburg (Bråthen 1982). In fact, in the dendrochronological record, we see many instances of timber that dates strongly against this chronology, including timbers from mid-16th century Aalborg (e.g. Sankelmarksgade, Daly 2016) and timber from Odense on Funen (Daly 2015). Timbers in Stirling Castle in Scotland from a building phase in 1592, also seem to come from this source (Crone 2008). Even elements of an altarpiece, dating to 1555, from as far away as Seville in Spain were built with west Swedish timber (Rodríguez-Trobajo, Domínguez-Delmás 2015). It seems clear that extensive trade in oak timber from western Swedish territory was taking place in the 16th and into the 17th century, and this timber is encountered in a wide range of contexts. That Amsterdam was a hub for timber from west Sweden eventually reaching Stockholm for the building of Vasa seems a very likely interpretation of these results.

3.3. Provenance Group 4

14Thirteen of the 14 timbers that are assigned to Group 4 are included in the tree-ring average for this group. The average is 244 years long and covers AD 1354‑1597. The distribution of correlation for this group is shown in fig. 3D. High correlations appear with chronologies from the eastern Baltic. The highest t-values are t = 8.64 with the so-called ‘Baltic 1’ art chronology and t = 12.18 with the ‘Baltic 3’ art chronology. On the map the placement of these art chronologies is conjectural. Other chronologies that Vasa’s group 4 match with (t = 6.83) include an oak site chronology from Vilnius (Rūtilė Pukienė pers. comm.). Vasa’s purchase records mention planks bought in Riga and Königsberg, so the combination of the historical evidence with the dendrochronological allows us take the evidence to the next level and suggest that the Baltic 3 art group might represent oaks from the east Baltic, rather than from Gdańsk and the Vistula. This is the first time we have come anywhere close to identifying the origin of one of these art-historical chronologies since the 1980s, when the Baltic origin of them was first discovered (Baillie et al. 1985; Eckstein et al. 1986).

3.4. Timber percentages from the four main sources

15To sum up the timber sources for Vasa we might look at the proportion of timber used in the ship from each group. This is presented in fig. 4 illustrated as percentages. The colours represent the colours assigned to each group in fig. 2 and the solid colouring shows the percentage of all timbers dated. However, many of the timbers that have been analysed from Vasa in this study are small pieces, like the sills that finished the gunports and the short ceilings inserted at the level of the gunports after the ports were cut through the hull. We might categorise these pieces as offcuts from larger timbers: so, in order to make a weighted count of the timber groups, I have taken these ‘­offcuts’ out of the count to achieve an indication of the volume of timber for each group shown using a lighter shade of the group colour codes. In the weighted count, we see that by far the greatest source of timber for Vasa is the two east Sweden groups, at 32.6% and 44% respectively. The timber from west Sweden, which came to Stockholm via Amsterdam, represents 6.4%, and the timber (all planks) from the eastern Baltic region just 7.1%.

Fig. 4: Diagram showing the percentage of timbers dated from Vasa from each group

Fig. 4: Diagram showing the percentage of timbers dated from Vasa from each group

The darker colours represent the percentage of all timbers analysed. The lighter colours show the percentage of full-length timbers in the different groups. The sills at the gunports and some other smaller timbers are, for this exercise, considered offcuts.

4. Conclusion

16The dendrochronological analysis of the Vasa timbers is an enormous undertaking, and this short article can only present the highlights of the research. Details of the various discoveries that have emerged are in preparation for publication, and the many aspects that I cover here will be expanded. The possibility to combine the details from written sources of timber procurement with the material analysis using dendroprovenance is rare. It allows us an insight into the logistics of assembling the resources for shipbuilding in a region dominated by forest in the early 17th century, demonstrating that the ways that this resource was attained were far from simple.

Acknowledgements

17
The TIMBER project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 677152), and Vasamuseet has kindly facilitated access to their collection. Thanks to Fred Hocker for many useful discussions during many trips to Stockholm, and to both Fred and Emma Hocker for their hospitality. Thanks also to Jim Hansson for giving me the opportunity to analyse timbers from the shipwreck in Stockholm harbour, subsequently identified as Scepter.

Haut de page

Bibliographie

Baillie M.G.L., Hillam J., Briffa K.R., Brown D.M.
1985 “Re-dating the English art-historical tree-ring chronologies,” Nature 315, p. 317‑319.

Bråthen A.
1982 “Dendrokronologisk serie från västra Sverige 813‑1975,” Rapport Riksantikvarieämbetet och Statens historiska museer 1, Stockholm.

Cederlund C. O.
1966 Stockholms skeppsgård 1605‑1640. Personalens struktur och organisation, SSHM internal report.

Crone B. A.
2008 “Dendrochronological analysis of the oak and pine timbers,” in Stirling Castle Palace; archaeological and historical research 2004–2008, Historic Scotland, http://sparc.scran.ac.uk/publications/level%20II/level2PublicationsMajor.html

Daly A.
2002 “Dendrokronologisk undersøgelse af tømmer fra Funder kirke, Århus amt,” Nationalmuseets Naturvidenskabelige Undersøgelser rapport 19, Copenhagen.

Daly A.
2013 “Dendrochronological analysis of timbers from three ships – BC01, BC03 & BC10, from Barcode, Oslo,” Dendro.dk report 233, Copenhagen.

Daly A.
2015 “Dendrokronologiske undersøgelse af tømmer fra Nørregade Fjernvarme, Asylgade, Odense, OBM6564,” Dendro.dk rapport 8, Copenhagen.

Daly A.
2016 “Dendrokronologisk undersøgelse af tømmer fra Sankelmarksgade, Aalborg, NJM6539,” Dendro.dk rapport 14, Copenhagen.

Daly A.
2017 “Dendrochronological analysis of timbers from a shipwreck and waterfront at Skeppsholmen, Stockholm,” Dendro.dk report 54, Copenhagen.

Daly A.
2018 “Dendrochronological analysis of additional timbers from a shipwreck at Skeppsholmen, Stockholm,” Dendro.dk report 5, Copenhagen.

Daly A., Streeton N. L. W.
2017 “Non-invasive dendrochronology of late-medieval objects in Oslo: refinement of a technique and discoveries,” Applied Physics A 123, 431, doi: 10.1007/s00339‑017‑1019-x

Daly A., Domínguez-Delmás M., van Duivenvoorde W.
Forthcoming “Batavia shipwreck timbers reveal the key to Dutch success in 17th-century world trade.”

Eckstein D., Ważny T., Bauch J., Klein P.
1986 “New evidence for the dendrochronological dating of Netherlandish paintings,” Nature 320, p. 465‑466.

Grupe G., von Carnap-Bornheim C., Söllner F.
2011 “Stable strontium isotope mapping for provenance studies in archaeology – different material, different signals?” Bulletin de la Société Suisse d’Anthropologie 17, p. 67‑76.

Hajj F., Poszwa A., Bouchez J., Guérold F.
2017 “Radiogenic and ‘stable’ strontium isotopes in provenance studies: A review and first results on archaeological wood from shipwrecks,” Journal of Archaeological Science 86, p. 24‑49.

Hillam J., Tyers I.
1995 “Reliability and repeatability in dendrochronological analysis: tests using the Fletcher archive of panel-painting data,” Archaeometry 37, p. 395‑405.

Hocker F.
2011 Vasa: A Swedish Warship, Stockholm, Medströms Bokförlag and National Maritime Museums in Sweden.

Lehner B., Grill G.
2013 “Global river hydrography and network routing: baseline data and new approaches to study the world’s large river systems,” Hydrological Processes 27.15, p. 2171–2186. (Data is available at www.hydrosheds.org.).

Rich S., Manning S. W., Degryse P., Vanhaecke F., Latruwe K., Van Lerberghe K.
2016 “To put a cedar ship in a bottle: Dendroprovenancing three ancient East Mediterranean watercraft with the 87Sr/86Sr isotope ratio,” Journal of Archaeological Science: Reports 9, p. 514‑521.

Rodríguez-Trobajo E., Domínguez-Delmás M.
2015 “Swedish oak, planks and panels: dendroarchaeological investigations on the 16th century Evangelistas altarpiece at Seville Cathedral (Spain),” Journal of Archaeological Science 54, p. 148‑161.

Wazny T.
2002 “Baltic timber in Western Europe – an exciting dendrochrono­logical question,” Dendrochronologia 20.3, p. 313‑320.

Haut de page

Table des illustrations

Titre Fig. 1: One of the gunports on Vasa
Légende Access to the end grain of hull planks and ceiling planks is possible. Furthermore, when the oak sill is removed (it can be seen lying on the deck) the cross sections of the sill and the frames can also be accessed.
Crédits (photography A. Daly)
URL http://journals.openedition.org/archaeonautica/docannexe/image/1919/img-1.jpg
Fichier image/jpeg, 2,0M
Titre Fig. 2: The tree-ring curves from each timber analysed from Vasa are compared with each other (correlation)
Légende The grey shades indicate the correlation values, but in order to simplify the illustration these values are not presented. Four main groups have been identified.
URL http://journals.openedition.org/archaeonautica/docannexe/image/1919/img-2.jpg
Fichier image/jpeg, 933k
Titre Fig. 3: The four maps show the correlation between the averages for the four Vasa groups and a range of oak master and site chronologies for Northern Europe
Légende A – Group 1 (blue), B – Group 2 (green), C – Group 3 (orange) and D – Group 4 (red). The river data is from Lehner & Grill (2013, www.hydrosheds.org, accessed 3rd March 2020).
URL http://journals.openedition.org/archaeonautica/docannexe/image/1919/img-3.jpg
Fichier image/jpeg, 2,0M
Titre Fig. 4: Diagram showing the percentage of timbers dated from Vasa from each group
Légende The darker colours represent the percentage of all timbers analysed. The lighter colours show the percentage of full-length timbers in the different groups. The sills at the gunports and some other smaller timbers are, for this exercise, considered offcuts.
URL http://journals.openedition.org/archaeonautica/docannexe/image/1919/img-4.jpg
Fichier image/jpeg, 241k
Haut de page

Pour citer cet article

Référence papier

Aoife Daly, « Timber supply for Vasa: new discoveries »Archaeonautica, 21 | 2021, 263-268.

Référence électronique

Aoife Daly, « Timber supply for Vasa: new discoveries »Archaeonautica [En ligne], 21 | 2021, mis en ligne le 01 juillet 2022, consulté le 28 mars 2024. URL : http://journals.openedition.org/archaeonautica/1919 ; DOI : https://doi.org/10.4000/archaeonautica.1919

Haut de page

Auteur

Aoife Daly

Associate Professor, Saxo Institute, University of Copenhagen

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-4.0

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