Introduction

Barczewko (ger. Alt Wartenburg) is a village located in northeastern Poland [Fig. 1]. Historically, it was an area of the Warmia episcopate established within the borders of the Teutonic Order, whose power at the Baltic Sea and northern territories of the Grand Duchy of Lithuania increased after the conquest of Prussia in 14th century1 (p. 151–169). At that time, in southern Prussia, the settlings of the wilderness areas continued to grow based on the so-called German law modelled on the Magdeburg legislation and its local modifications. Within the Teutonic state, parallel settlement activities were carried out by bishoprics, among which the Warmia region was economically dominant. Its rulers—the bishops, considering themselves subordinate only to Rome, developed an original system, economic power and autonomy lasting until the first partition of Poland in 1772.

Fig. 1: Barczewko, warmińsko-mazurskie voivodship, Poland.
figure 1

Localization of the site. The other marked sites represent the original sources of the reference chronologies used in this study. (by A. Koperkiewicz).

Full size image

A witness of the era, the Teutonic chronicler Peter of Dusburg, reports the construction of the Wartenborg castrum in 1325 on the Pisa River, which natural defensive qualities such as bends were probably used for this purpose. It was one of the last initiatives of Bishop Eberhard of Nysa. On his behalf, logistics were handled by a special official, bringing in settlers from the bishop’s native lands of Silesia2. A lot of artifacts found in Barczewko (military, tools, cargo seals from Venice) show an analogy in the finds from Wrocław3 (p. 491-492), and indicate that the inhabitants could have come from Silesia Region. These people successfully transferred an urban culture and lifestyle from their homeland to a foreign land, and they most likely had some contact with the indigenous population1 (p. 134–135);4. Only after the area had been reconnoitered, identified, and secured (including the construction of a watchtower), could the process of bringing in the remaining settlers begin. The local Prussian population, familiar with the area, played an important role in this process. We can observe the same type of process (conquest, settling, population displacement) in parallel time period on the British Isles5, (41);6, (p. 188-89). Landscaping of an untouched area, with no previous settlements, presumably started between 1325 and 1354. It was a time of an increased demand for the basic building material—wood. Historical documents mention a parish priest, which prompted researchers to search for the remains of the church, which—next to housing—was the most important investment in Christianized areas.

From another testimony we learn the date and circumstances of the destruction of the settlement. The chronicler Wigand of Marburg records that in 1354 the Lithuanian army launched a surprise attack and burnt Wartenburg. This seems to be nothing exceptional at that time, yet the tragedy of the inhabitants is still mentioned 100 years after these events in the chronicle of Jan Długosz from the middle of the 15th century. The number of attackers was supposedly so huge that the knights of the Teutonic Order failed to even attempt a battle. Perhaps the death of the defenceless inhabitants was the reason why the drama of this site was encircled by a special taboo and remained intact for centuries. The town was rebuilt in 1364, in a new location, and the bishop fenced it with towers and ramparts stronger than the previous ones. It is also obvious from the testimony that the first ones were too weak.

Almost seven centuries after the recorded events, a stopped in time snapshot of the last day in the life of the young colony was unveiled. The destruction of Barczewko was recorded in archaeological layers with traces of a huge conflagration. The starting point for the research was sparse but reliable historical sources. The analysis of aerial photographs and geophysical anomalies left no doubt that the hill hid preserved relics of an organized structure of a medieval town7. In 2014, research works on a medieval lost city began as part of a project that was somewhat exaggeratedly called ‘the Warmian Pompeii’. The results were so promising that the works continued in subsequent seasons8,9,10,11,12 and led to another project entitled ‘Barczewko. Historia ogniem pisana’ (eng. Barczewko. History written by fire) which is currently being conducted.

The carbonized but fully preserved construction elements (pillars, joists, planks) provided a unique set of dendrological samples. The preliminary analyses carried out in 2014–2016 showed how much analytical potential the excavated material holds. The dating of the charcoals analyzed at that time oscillated around the first decades of the fourteenth century13,14. Complementing these works with interesting and certain research results acquired in subsequent seasons, we tried to answer the question of whether the construction history of the settlement could be traced and whether dendrochronology as a scientific discipline would confirm the historically documented existence of Barczewko in such a narrow, only two decades long period between 1330 and 1354. The potential and high applicability of dendrochronological analysis for the historical sciences is presented in this article, showing not only absolute dating of the wood and the developed local chronology, but also discussing unique anatomical features of the timber, its origin and species identification.

Methods

Archaeological research

Barczewko’s architecture was exclusively wooden. During the fire, some of the cellars and their contents were pressed down by the weight of the walls of the upper parts of the buildings. Thus, most of the artefacts were ‘trapped’—grain containers, traces of barrels, bowls, pots, agricultural tools, fishing implements, and some of them were malformed by the impact of high temperatures [Fig. 2].

Fig. 2: The relic of the burnt building ‘smithy’.
figure 2

(by A. Koperkiewicz).

Full size image

Archaeological discoveries were being located in subsequent seasons with the use of geophysical anomaly maps. The following were distinguished: remains of cellars, mercatorio complex, post pits, farm pits and hearths, relics of fortifications, natural and false anomalies. Among the rubble with several coins, an abundance of militaria was found, the vast majority of which were crossbow boltheads. The remains left after the conflagration covered basement pits with an accumulation of in situ artefacts, as well as the remains of human and animal victims [Fig. 3]. This is the classic ‘Pompeian case’ in archaeology, where a short-term historic event, well-settled in time, sudden and tragic, becomes a repository of information about the era.

Fig. 3: Barczewko, selection of artefacts.
figure 3

Clockwise: a pottery in building no. 11, b copper cauldron with vessels and tools, c victim’s skeleton in building no. 9, d spur in building no. 16, e coin deposit in building no. 15 (by A. Koperkiewicz and T. Waszczuk).

Full size image

From the perspective of both archaeological and dendrochronological studies, the cellar relics were the most significant, due to the very large number of carbonized structural elements found inside. The research began with those cellars, and later focused in the central part of the site uncovering further constructions around the market square—the mercatorio. The church and the cemetery were located on the NE side.

Corner and center posts [Fig. 4] were inserted ~0.5–1 m deep into a square-shaped trench; the basement walls were made of wooden planks ca. 30 cm wide and 3 cm thick. No remains of braces were found and the connecting details, due to significant carbonization, were illegible. It seems that these were simpler columnar structures with braced transoms.

Fig. 4: A post supporting the framing of the basement wall embedded in the ground.
figure 4

Sampling material was best preserved in the lower parts (by F. Popek).

Full size image

Early in the 14th century, timber-frame constructions had been already fully developed, however the excavated basement walls were formed by framing cladding with thick beams, and exterior soil providing adequate isolation, as evidenced by the lumps of burnt clay, with imprints of wood and sticks on the surface [Fig. 5].

Fig. 5
figure 5

Fragment of ‘polepa’, the burnt clay filling the wall structure with visible admixture of organics (grass, sticks) and imprints of wooden elements (by W. Brillowski).

Full size image

Among the relics, at least two structures were similar to each other, but different from typical cellars – they were much smaller, and there were no remnants of the columnar structure typical of cellars. The layers were strongly compressed and contained a lot of typical settlement material (pottery, animal bones, fish bones, but also coin finds). In both cases, a rectangular outline of a burnt wood layer with traces of a circular hearth in the north-western corner was formed at the bottom of the pit. These may have been simple dugouts prior to the proper development, but such a purpose does not exclude the possibility that they may have later functioned as farming buildings.

On the western line of the market frontage two twin structures were exposed. A concentration of about 160 crossbow bolt arrowheads was discovered in the central part of the entrance, along with ornamental lock fittings, and silver clasps of the Ave Maria type, or so-called ‘engagement clasps’, which are unique on the site. To the rarer finds, we can add a fragment of a spur and a conical lead weight.

The objects were perfectly legible on geomagnetic visualisation [Fig. 6]. After removing the arable layer, it became clear that the basement of one building has an unusual zigzag entrance. Relics reached a depth of about 2 m. Exceptionally legible were the structures in the basement of building no. 11 (obj. 594), where charcoals of the horizontal foundation with columns at the corners were preserved. The identification of the western cellars (no. 10 and 11) was crucial for establishing the house unit and the building measure. Between the northern and middle lines, on the west side of the basement of building no. 11, a row of post-holes was exposed, arranged in such a way that the free space was exactly opposite the entrance to the basement of building no.11 [Fig. 7].

Fig. 6: Planigraphy of the distinguished archaeological relics (black) on the map of geophysical anomalies (red and blue).
figure 6

The example of the reeves' house (buildings nos. 10 and 11): a plant features, b magnetic features, c archaeological negatives, d explored relics. o/B - numbering of objects /buildings, e.g. B4 and B9 – cellars with sceletons of victims, B6 - bath house, B8 - fisherman's house, B3 - farmer's house, B10 and B11 - reeve's houses, B13 - smithy, I - gate, II - communication route, III - market square, IV - reeve's parcels, V - craftsmen's trade complex, VI - cemetery, VII – church. Prepared by A. Koperkiewicz, map by P. Wroniecki.

Full size image
Fig. 7
figure 7

Planigraphy of the exposed relics of the spatial layout of the town with dendrochronological dates for particular objects obtained by Uwe Heussner13, Tomasz Ważny14 and Anna Elzanowska (Table 1). Prepared by A. Koperkiewicz, map by M. Marczewski.

Full size image

The distances between the central lines of the pits were ~8.7 m, and the length of the surviving line was ~17.5 m. The measured distances fit quite precisely with the so called ‘Chełm bar pattern’—a medieval measurement pattern equal to 4.32 m and used in the 13th century to establish cities on the so-called raw root, i.e. from scratch, on previously uninhabited land. [Fig. 8]. The buildings are a model example of an ‘Altstadt’, a town from the 2nd half of the 13th century, where we find many analogous construction solutions15.

Fig. 8
figure 8

The plan of the uncovered relics of buildings Nos. 10 and 11, together with the basement and the farm section marked by a sequence of post pits. The base of the module is on a rectangular plan measuring 8.7 x 13 m. Prepared by P. Dudek, A. Napierała and A. Koperkiewicz.

Full size image

A large number of wooden building elements were uncovered in the eastern line of the market square in the cellars of buildings no. 4 and 9. Buildings no. 5 and 14 neighbored the boundary of the cemetery. Due to the peculiar state of preservation of the relics in cellar no. 5, it was assumed that the northern market building was then under construction. A very large number of carbonized wall structures and a cellar ceiling pressed down by clay were preserved, with an almost complete absence of movable monuments [Fig. 9]. Compared to other towns in the Teutonic state, the market foundation was small (~43 m x 48 m), which is characteristic of later endowments and episcopal locations.

Fig. 9: Burnt relics of building No. 5 in the northern part.
figure 9

Wooden elements of the wall and ceiling structures along with burnt clay (by F. Popek).

Full size image

Tree-ring analysis

Visual inspection of the charcoal samples was carried out to assess the state of their preservation, suitability for study and selection of appropriate elements. Selection followed the site diagrams and was also guided by samples whose status was determined as representative of the site.

153 samples were received for the purpose of current dendrochronological analysis. They were found in varying state of preservation, and 54% of them—82 samples—were qualified for dendrochronological studies. For material that was heavily deformed, split, preserved in powdered form or in an advanced stage of fungal decay, further analyses were limited to species identification.

Identification of the tree species was performed with the use of microscopic methods. Surface revealed at cross-section, tangential and radial sections was observed in selected samples. The charcoal pieces were examined under a DELTA Optical microscope at the magnifications of 50x, 100x and 200x. The obtained microscopic image was compared with wood atlases and reference materials16.

The sample preparation, measurements, and further analyses followed traditional dendrochronological methods17,18. Suitable surfaces were marked out and prepared at the cross-section of samples. The tree-ring widths were then measured on a LINTAB-type measuring stage. The measurement accuracy was 0.01 mm.

Tree-ring series were synchronized against available reference chronologies from the broader study area19,20. This was done separately for Pinus sylvestris and Quercus sp. as other species were not qualified for the dendrochronological analysis. Both visual and statistical cross-dating was carried out. For statistical cross-dating the following parameters were used: Gleichlaeufigkeit (Glk)—which represents the percentage of year-to-year changes in ring widths which are the same (both increasing or both decreasing) between the ring-width sequence in a section and the reference chronology21, t–value Hollstein (TVH)22 which are sensitive to extreme values, such as marker years. TVH values above 4.0 were considered statistically significant and taken into account for further visual comparison. We also analyzed the Cross-Dating Index (CDI), which combines all correlation measures23 and is significant for CDI > 25. The common interval expressed by the number of overlapping years (OVL) was also considered. After obtaining dates for the individual series, we developed a mean chronology and repeated the visual and statistical cross-dating process using the same reference chronologies. TSAPWin23 and DENDRO for Windows 1024 were used for cross-dating and visualization of the results.

It should be highlighted how valuable the Barczewko site has been to many researchers. Some dendrochronological studies were undertaken by other specialists within the projects in 2014 and 2016 and based on the samples available at that time. However, they were never published, although the obtained results showed the potential of the site13,14. Only deeper excavations conducted in the following years made it possible to have a full overview at the Barczewko Old Town, and the dendrochronological analysis was carried out in thorough comparison with archival sources.

Results

Species identification

The samples represented eight recognized tree species [Fig. 10]: Quercus sp.—92 samples, Pinus sylvestris—33 samples, Fraxinus excelsior—2 samples, Tilia sp.—11 samples, Betula sp.—6 samples, Ulmus sp.—1 sample, Platanus sp.—1 sample, Fagus sp.—1 sample. In a few samples, detailed identification was not possible due to the state of their preservation (powdered material) – those samples were identified in general as unspecified deciduous trees—6 samples.

Fig. 10
figure 10

Bar diagram showing the recognized tree species and genera. (by A. Elzanowska).

Full size image

Tree-ring chronologies

40 samples were dated and this represents as much as 49% of the selected material suitable for dendrochronological analysis [Table 1]. It is worth mentioning, that 10 samples could be dated with a very high degree of precision, and among them—four were dated with accuracy 1 year due to the presence of sapwood boundary. The rings in oak samples were very narrow [Fig. 11], showing the slow growth of the trees in the forest spreading back then within the territories of today’s Poland, Belarus or Lithuania. Oak is an exception nowadays in wooden constructions in Warmia Region, and since the 15th century almost everything has been built from pines (Pinus sylvestris). The well-preserved charcoal remnants of Barczewko still reflect the original forest vegetation and high quality of the timber, resilient against mechanical pressure and other external factors.

Fig. 11
figure 11

Microscopic image presenting very narrow tree rings in an oak sample Brczw30a. (photo by A. Elzanowska).

Full size image
Table 1 Information about the dated Barczewko samples
Full size table

In many cases, it was not possible to determine the precise felling date due to the absence of the outermost rings (which determine the accuracy of the dating, sometimes even with annual precision).

As a result of the study—after having obtained precise dates for particular elements of the site–the BrczwQEW local chronology covering the time span of 253 years (1081-1333) was created from the dated oak samples. Oak were dominant species found during all studies conducted on the Barczewko material, and therefore were selected for the development of the chronology. The characteristics of BrczwQEW chronology is presented in Table 2.

Table 2 Characteristics of the developed BrczwQEW oak chronology
Full size table

The master BrczwQEW chronology created for the Barczewko site provides a unique matrix for further dating of artefacts from this still dendrochronologically abandoned region. It is one of the longest oak chronologies obtained exclusively from charcoals and absolutely dated by means of classic dendrochronology. The results of the statistical comparison of BrczwQEW and reference chronologies is presented in Table 3 and Fig. 12.

Fig. 12: Visual cross-matching of the developed BrczwQEW oak chronology (blue colour) and the reference P671001M oak chronology (purple colour) showing excellent growth-pattern correlation.
figure 12

Only the overlapping period of the reference chronology is presented here. (by A. Elzanowska).

Full size image
Table 3 Statistical comparison of BrczwQEW chronology and reference chronologies
Full size table

Moreover, the BrczwQEW chronology corresponds well with 2021BLT1 chronology covering time span 1143–1626, which is one of the chronologies developed for the so called Baltic Oak, a term which refers to the oak timber trade in 13th–17th c., when oaks from the area of nowadays Poland, Lithuania or Belarus were exported to Flanders, northern Belgium or even England, mainly for the purpose of panelmaking25 or even construction of ships and docks. This indicates that the oak timber was felled and exported to Western Europe, also from the area of primary forests located near Barczewko, as the Galindian Forest was part of a bigger complex spreading up to current coastal Lithuania. The wood was transported over the distance of even thousands of kilometres26 (p. 509–513) through the agency of the German Hanse and the Teutonic Order up to the middle of 17th c27.

Discussion

In archaeological material of this type (charcoals), significant damage of samples due to heat and their mechanical weakness is to be expected. In many cases, the number of missing outer rings cannot be estimated, but each sample and its ring pattern was analyzed individually to obtain the most accurate dating. Interpretation was based on the dating distribution of the entire, large group of samples.

The analyzed elements from Barczewko originate from the end of the 13th century and the turn of the 13th/14th century at the earliest. Taking into account the outermost and burnt layer, which has crumbled and fallen off to varying degrees (eroded), we can estimate the 1330 s for at least four youngest samples (samples Brczw09a, Brczw15a, Brczw20a, Brczw28a). What definitely needs outlining, the only accurately dated piece (sample Brczw28a) was made from a tree felled between 1335 and 1337. The oldest pieces (samples Brczw18a and Brczw19) date to ca. 1288 and ca. 1273, respectively, and most likely they come from reused timber or from the felling of previously dead trees.

It is worth mentioning that the samples well dated to the late 13th century or early 14th century are problematic in interpretation. The following possibilities may be considered:

a/ the use of timber fallen or felled earlier, then stored in already existing towns in Warmia and transported for the construction of Barczewko;

b/ secondary use of wood from existing nearby Prussian settlements or watchtowers on the Pisa River, known to have existed in the early Middle Ages (so called Olsztyn group);

c/ the retrieval of timber from demolition of pre-existing episcopal towns (Orneta, Dobre Miasto, Lidzbark Warmiński).

The last aspect is very interesting and has never been taken into account before. It is known that Bishop Eberhard assigned the task of organising the town to his clerk. It is not excluded that when conducting a settlement campaign, logistical preparations were made in already fortified centres which had functioned since the beginning of the 14th century. This way, the new group of settlers was not left alone and such an option seems plausible due to the specificity of the episcopal jurisdiction.

Given its time span, the tree-ring structure, and the high correlation with oak chronologies from northern Poland, especially nearby Dąbrówno, Barczewko’s timber is of local origin and most likely comes from the Galindian Forest (pol. Puszcza Galindzka/Wielka Puszcza, ger. die Grosse Wildnis) [Fig. 13] or a nearby forest complex, although at this time the area had already been subjected to high deforestation28. We can assume that wood was a highly demanded material for the construction of a new settlement, and it was reasonable to source it in the closest forest complex available.

Fig. 13: Examples of unsuccessful town locations shown against the backdrop of Prussian tribal territories and the outskirts of the Galindian Forest: Barczewko (Alt Wartenburg) and Old Town (Castle Mountain) near Dąbrówno.
figure 13

The area of historical Warmia, which was governed by bishops during the time of the Teutonic Order, is marked in orange. (prepared by A. Koperkiewicz, map by M. Sobiech).

Full size image

The finding of Platanus sp. in one sample is a unique discovery in this part of Europe, as the natural range of this species covers the Mediterranean and Black Sea Basin29 (p. 224). This sample comes from a barrel element (bottom) found in the object nr 954 (NE part of cellar nr 16, trench 38/2024) and it was discovered among other interesting artifacts, which altogether form an intriguing archaeological context. As Platanus sp. represents some anatomical characteristics similar to Fagus sp., a thorough study on the features was conducted. This analysis revealed diffused porous vessels in the number of ~110 per square mm and wide rays ~10-seriate, which definitely allowed us distinguish the sample from beech (Fagus sp.). Additionally, to exclude the possibility of having a later (younger) finding, it should be outlined that the barrel element was found surrounded by other objects/fragments of objects that assure us about the undisturbed context of the place (e.g. traces of glass and some very fragile vessel with ornaments, some spurs, many coins, shovels and Venetian cargo seals) [Fig. 14]. A fragment of the door (with a piece of board jointing) illustrates how well preserved were the charred findings of Barczewko [Fig. 15].

Fig. 14: The relic of burned building no. 16 ‘praetor’s house’.
figure 14

On the bottom of the basement there are scattered ceramic vessels, in the central part there is an imprint in the clay of an item from a wooden ceiling structure, on the right there is a charred bottom of a barrel made of plane tree. Perhaps this should be linked to the discovery of Venetian seals nearby. (by A. Koperkiewicz).

Full size image
Fig. 15
figure 15

The preserved element of the entrance door in the basement of building no. 16. (by A. Koperkiewicz).

Full size image

Taking into account the above, as well as the origin of the sample or the depth at which it was excavated, it can be concluded that the barrel is excellent evidence of trade between Prussia and the Mediterranean or Black Sea countries. The wood of plane tree was used among other for fruit and vegetable boxes and dry substance barrels30, so this finding may be a testimony to how far north goods were exported from southern Europe in the 13th–14th centuries, and definitely deserves a separate thorough study.

Regarding the animal and human remnants found, it is worth mentioning that, in the context of historical data, the question of victims has intrigued researchers from the very beginning. The fact that there were located such remants as an infant skeleton (building no. 4) and a woman skeleton (building no. 9) is the evidence that the site remained not landscaped after the battle and fire. Possibly there are other similar locations in the unexplored neighborhood. However, no answer was found to the question of what had happened to the victims in general. There must have been much more of them. They were certainly not buried at a cemetery where all the dead are placed in a certain order and according to ritual. Is there a mass burial site somewhere else, or perhaps the bodies of the victims were entirely burnt? One possible scenario is that there was no longer a person to bury the victims because most or all of them died. It is also possible that the survivors, or the new generation rebuilding the city elsewhere, treated the first location as an untouchable taboo area—as a huge graveyard. This is the only way to explain the phenomenon of valuable objects being left in the cellars, the hillside being abandoned and, above all, the lack of Christian burial required by the Church.