Radiocarbon dating of the Church of St. Margaret of Antioch in Kopčany (Slovakia): International consortium results

Abstract

An international consortium of radiocarbon laboratories has established the origin of the Church of St. Margaret of Antioch in Kopčany (Slovakia), because its age was not well known from previous investigations. In total, 13 samples of charcoal, wood, mortar, and plaster were analyzed. The ¹⁴C results obtained from the different laboratories, as well as between the different sample types, were in good agreement. Resulting the final ¹⁴C calibrated age of the Church, based on dating a single piece of a wooden levelling rod is 774–884 AD (95.4% confidence level), which is in very good agreement with Bayesian modeling result based on dating of wood, charcoal and mortar samples (788–884 AD, 95.4% confidence level). The probability distribution from OxCal calibration shows that 79% of the probability distribution lies in the period before 863 AD, implying that the Church could have been constructed before the arrival of Constantine (St. Cyril) and St. Methodius to Great Moravia. If we take as the terminus post quem the documented date of consecration of the church in Nitrava (828 AD), the Bayesian modeling suggests the age of the Church in the range of 837–884 AD (95.4% confidence level). Although the ¹⁴C results have very good precision, the specific plateau shape of the calibration curve in this period caused a wide range of the calibrated age. The Church represents, together with the St. George’s Rotunda in Nitrianska Blatnica, probably the oldest standing purpose-built Christian church in the eastern part of Central Europe.

Introduction

The Church of St. Margaret of Antioch made of stones and mortar (Figure 1) is located close to the village of Kopčany on the bank of the Morava River in southwestern Slovakia (Figure 2). Although it passed several reconstructions over the centuries, its shape did not change much. Recently, a restoration and reconstruction of the roof was carried out to give the Church its original look, including calcareous plaster, which should protect the walls against further deterioration. Therefore, this was a chance to collect samples from the external and partly also from internal walls of the Church and submit them for radiocarbon dating. Because of the archaeological and historical significance of the Church and to exclude possible discussions on results obtained only by a single national radiocarbon laboratory, an international consortium of radiocarbon laboratories from neighbouring countries (Austria, Hungary and Czech Republic), one from Switzerland and two from USA, was organized.

Figure 1. Church of St. Margaret of Antioch located close to the village of Kopčany in Slovakia (pre-reconstruction and post-reconstruction views).

Figure 2. Map of western Slovakia with the location of the Church of St. Margaret of Antioch in Kopčany. The capital of Nitrava Principality and the location of the St. George Rotunda in Nitrianska Blatnica are also shown.

As in our previous paper in this series (Povinec et al. 2021), we have already described in detail the historical background, therefore, in this paper we just mention only a few historical facts. It is expected that in the period of the first half of the ninth century Christianity was accepted in some of the Central Europe countries (Goldberg 2006). According to written historical documents (Anonymous 870), Salzburg Archbishop Adalram consecrated already in 828 AD a church in Nitrava (presently Nitra), built by Prince Pribina, a ruler of the Principality of Nitrava, which was situated in the western part of the present Slovakia (Figure 2). He was the first ruler of Slavic origin who built a Christian church in central Europe. Unfortunately, the existence of this church is documented only in writing, since (after several reconstructions) it probably became a part of the Cathedral Basilica of St. Emmeram in Nitra. In 833 AD the Principality of Nitrava together with the Principality of Moravia formed a joint nucleus, which was later known as Great Moravia (Kingdom of the Slavs), which under Rastislav and especially under Svatopluk, was one of the largest kingdoms in Central Europe (Homza 2018).

Christianity in Central Europe considerably expanded after the mission of Constantine (St. Cyril) and St. Methodius, who arrived in Great Moravia in 863 AD. They had great impact on further religious and social developments in Great Moravia, also due to the creation of the Glagolitic alphabet and translation of the New Testament and Christian liturgy to the Slavonic language. The approval of the Slavonic liturgy in 868 AD by Pope Adrian II as the fourth official liturgical language (after Hebrew, Greek and Latin) (Terry and Gallagher 2017) was also a great event that helped the spread of Christianity in Central and Eastern Europe (it took eleven hundred years when after the II Vatican Council in 1963, Latin was replaced as the liturgical language by national languages). Later, in 895 AD the Principality of Hungary was established in the Carpathian Basin, and in 1000 AD King Stephen I formed the Kingdom of Hungary. An important part of these new developments was also the progression of Christianization from the west to the east, with competing interests between the centres of Rome and Constantinople (Terry and Gallagher 2017). All these historical events had a great impact on the cultural and social developments of Slavs during and after the Great Moravia times, when many Christian churches were constructed, and Christianity further spread in Central and Eastern Europe (Homza 2018; Spiesz and Caplovic 2006).

It is also interesting to mention that just over the Morava river from the Kopčany Church (on the right bank of the river, currently connected with a bridge for pedestrians) is the Mikulčice settlement, which was one of the main centres of the Great Moravian Empire. Excavations discovered remnants of twelve churches, a palace, and more than 2500 graves, but unfortunately no standing building was preserved. The Great Moravia lasted from 833 to 907 (of which the Principality of Nitrava was a part), and during this time Christianity in Central Europe considerably expanded, especially after the arrival of St. Cyril and St. Methodius to Great Moravia (Kouřil 2014).

Past archaeological and restoration investigations

The Church of St. Margaret of Antioch in Kopčany, due to its possible association with Great Moravia has been of great interest to historians and archaeologists. Several studies have been published in the past. Here we mention at least the most recent ones which were focusing on estimation of the Church age (Bahyl et al. 2013; Baxa et al. 2015; Botek et al. 2015a, 2015b). A lot of works connected with restoration of the Church were not well documented, therefore, its origin was not well known. Before 1994 it was believed that the Church was of baroque origin; later its origin was shifted to the gothic period. However, its proximity to the Great Moravian centre at Mikulčice (just across the Morava River) suggests that it may have originated in the 9th–10th century.

The Church has a typical floor plan with one nave, about 4.8 m long and 3.8 m wide; the back side is 1.8 m long and 2.3 m wide (inside dimensions). The height of the walls is 5.6 m from the original level of the terrain. The walls were built from local split stones, forming a row masonry of anti-conical arrangement of about 60–74 cm thick. This type of construction made the walls very stable even without a roof, which could be missing sometimes in the past. However, such arrangements caused difficulties when opening the wall to collect mortar samples from its interior.

Archeological and architectural-historical investigations of the Church were carried out mainly at the end of the 20th century (Baxa 2000; Baxa et al. 2004, 2005; Botek 2007, 2010; Puškár and Mýtnik 2006, 2009; Sabadošová and Havlík 1998, 1999), which then continued at the beginning of the 21th century (Baxa et al. 2015; Botek 2013; Botek et al. 2015a, 2015b; Puškár and Mýtnik 2009). Archaeological research also included immediate surroundings and a cemetery (also located inside the Church), and brought new discoveries, suggesting that the Church is much older than previously thought. Baxa et al. (2005), summarizing the results of archaeological research and on the basis of the relationship between the Church and the graves of the church cemetery (including the dating of findings from the graves), suggested its first building phase to the 9th and the beginning of the 10th century. The jewelry found in the graves in the nave of the church was characteristic of Great Moravian, and they were also dated to the 9th century and the beginning of the 10th century as well.

The one-nave church with a straight closure (or the rectangular apse) presbytery, supported by archaeological findings, may suggest that the Kopčany church could belong to the Carolingian cultural circle, influenced by the dominant position of the Salzburg archbishop in the region. Almost identical floor plans could be found in the church of St. Kilian in Höxter (780–800), of St. Justus, Laurentius and Bartholomew in Flums (around 800), of St. James and Martin in Rauris (9th century), churches in Sudburg (9th century), in the church of St. Laurentius in Winterthur (9th–10th century), and in some others (Baxa et al. 2005). The layout of the Church in Kopčany, when compared with Great Moravian sacral architecture (e.g., in Mikulčice) is, however, atypical. Therefore, further archaeological research is needed to determine a function of the Church, because it is quite possible that it could be a monastery church.

The results of direct dendrochronological and radiocarbon dating of a wood sample found in the wall of the Kopčany church were reported by Bahyl et al. (2013). According to the dendrochronological analysis, the Church was built in the middle of the tenth century. Radiocarbon dating gave the ¹⁴C conventional age of 1025 ± 60 BP and the calibrated age of 890–1161 AD (95.4% confidence level) (calculated with the IntCal20 calibration curve of Reimer et al. 2020). As the ¹⁴C measurements were carried out by a traditional radiometric method (a gas counting), the uncertainty of the conventional age is by about a factor of 3 higher than that obtained by present accelerator mass spectrometry (AMS).

Several reconstruction works were carried out during the history of the Church (Sabadošová and Havlík 1998, 1999, 2008; Botek et al. 2015a, 2015b), to mention at least a few of them. The first changes were associated with closing windows on the west and east sides, changes in the height of the floor, making new plasters, and probably also a vestibule was pulled down (Drahošová and Vančo 1996). During the 10th–12th century two small windows on the north side of original design were closed, and new plasters were added. In the 14th–15th century, originally small windows on the southern wall were replaced with larger Gothic windows, and further renovations of plasters were carried out. In the 15th–16th century probably the most important renovation was setting of the broken portal in the western wall, and new plasters covered with paintings were made (unfortunately only torsos of them were preserved). During the 17th–19th century several changes were made with reconstruction of windows, but probably the most important change were again new plasters with paintings. The last renovations were done in 1926 when a new lime-cement plaster was added to the interior and a thick layer of a coarse-grained structured plaster covered the exterior of the Church.

The restorations and reconstructions carried out recently have had the main idea of restoring the expression to its primary form, when after many years its external walls are again covered with single-layer lime plasters. Reconstruction of the exterior plasters was carried out in the structure and shape according to the preserved torsos of the primary plaster. The shields and roof of the church were reconstructed in the last known form from the Baroque stage of its development. The church has been plastered since its inception; thanks to the roof and plaster, the walls of the church have been preserved for more than 1000 years. The restoration works carried out on the interior and exterior of the Church (Kekeši 2014; Lupták 2021, 2022; Puškár and Mýtnik 2009) helped restore its original look and preserve it for future generations.

Fortunately, during the restoration works we were allowed to collect samples of charcoal and wood that were found in the walls, which together with samples of mortar and plaster collected from the inner and outer walls, were used to determine the age of the Church by radiocarbon dating. Because of the archaeological and historical significance of the Church, not only in respect to Slovakia but also to Central and Eastern Europe, it was decided that an international consortium, consisting of independent world-renowned radiocarbon laboratories, will be formed to carry out the dating work.

Samples

The sampling was carried out mainly on the exterior of the church during the restoration of the torsos of the primary plasters, as the interior walls have already been covered with plaster and white paint. In the course of the restoration activities, eight small samples of charcoal and one sample of wood were found on the stone walls after the surface mortar material was removed. Three mortar samples were also taken from the exterior walls. From the interior of the church, only one plaster sample was collected, not to disturb the restored plasters and paintings on the walls (Figure 3).

Figure 3. Inside the reconstructed church.

In total, 13 samples were analyzed: 8 charcoals, 1 wood, 3 mortars, and 1 interior plaster. The collected samples were divided and sent to at least two consortium laboratories for separate radiocarbon analysis. AMS has been used in all laboratories for ¹⁴C measurements.

Wood sample

Sample W-1: A branch piece (2 cm in diameter and 21 cm long) was found in a mortar hole (4 cm in diameter) in the southern wall, 280 cm above the ground, 60 cm from the right corner (Figure 4). The wood is expected to be a remnant of the levelling rod used during the construction of the Church. The dated wood piece originated from the centre of the branch, which could have 10–15 annual rings, compared to five rings that could be found in the collected sample of 2 cm in diameter. Actually, we have been very lucky that we found this wood piece, as it could be the best indicator of the age of the Church, as with dating of charcoal samples, we always have larger uncertainty because we cannot determine a place of the charcoal origin inside the wood.

Figure 4. A branch (W-1) found in a hole in the southern wall about 2.8 m above the ground.

Charcoal samples

Sample C-1: Small pieces of charcoal found in the exterior mortar of the southern wall at a height of 160 cm above the ground, 100 cm from the left corner (Figure 5).

Sample C-2: Small pieces of charcoal found in the exterior mortar of the southern wall at a height of 170 cm above the ground, 120 cm from the left corner.

Sample C-3: Small pieces of charcoal found in the exterior mortar of the southern wall at a height of 70 cm above the ground, 170 cm from the left corner.

Sample C-4: Small pieces of charcoal found in the exterior mortar in the centre of the northern wall, 150 cm above the ground.

Sample C-5: A sample found in the interior mortar of the northern wall, 180 cm above the ground, 150 cm from the apse.

Sample C-6: A sample found in the outside mortar in the northern part of the apse wall, 110 cm above the ground.

Sample C-7: A sample found in the outside mortar in the eastern part of the apse wall, 150 cm above the ground

Sample C-9: A sample found in the interior mortar of the western wall, 215 cm above the ground, 115 cm right from the entrance.

Figure 5. A small piece of charcoal (C-1) found in the exterior mortar of the southern wall.

Mortar samples

Sample M-1: A sample taken from the outside mortar of the south wall, 80 cm above the ground, 320 cm from the left corner.

Sample M-2: A sample taken from the outside mortar in the centre of the eastern part of the apse wall, 100 cm above the ground.

Sample M-3: A sample taken from the outside mortar in the northern part of the apse wall, 80 cm above the ground.

Plaster samples

Sample P-1: A sample of coarse-grained interior plaster was taken from the western wall, 215 cm above the ground, 115 cm right from the entrance. The C-9 charcoal sample was buried in mortar just below the plaster cover.

Methods

Radiocarbon analyses

The samples analyzed were divided and sent to at least two radiocarbon laboratories for independent ¹⁴C analyses. Each laboratory used its own sample pretreatment and measurement methods. All laboratories used AMS for ¹⁴C measurements. Carbon dioxide prepared from fossil carbon sources was used as a background sample. The measured activities and uncertainties were calculated following the Stuiver and Polach (1977) convention.

The Athens Laboratory at the University of Georgia

Standard procedures were applied for the pretreatment of the analyzed wood and charcoal samples using the acid-base-acid (ABA) method. The samples were then combusted, and the resulting carbon dioxide was cryogenically purified from other reaction products and catalytically converted to graphite using the method of Vogel et al. (1984). The ¹⁴C/¹³C graphite ratios were measured using the CAIS 0.5 MeV accelerator mass spectrometer (NEC, Middleton, USA). The sample ratios were compared to the ratio measured from the Oxalic Acid I Standard (NBS SRM 4990). The ¹³C/¹²C ratios of the samples were measured separately using a stable isotope ratio mass spectrometer and expressed as δ¹³C with respect to PDB (uncertainties were less than 0.1‰). The obtained ¹⁴C results were corrected for isotope fractionation.

Bratislava Laboratory at the Comenius University

The Bratislava group prepared graphite targets for AMS measurements, which were later carried out at the Athens Laboratory. Standard procedures were used for the pretreatment of the analyzed wood and charcoal samples. The charcoal samples, after visual control for possible impurities, were treated using the ABA method (Gupta and Polach 1985): a treatment in 4% HCl at 70°C, followed by treatment in 4% NaOH at 70°C, and finally treatment in 4% HCl at 70°C. For the wood sample, an additional bleaching step was added to separate cellulose from the sample. The pretreated samples were then combusted at 950°C in a quartz tube filled with pure oxygen. The carbon dioxide obtained after purification was then used for the preparation of graphite targets using the hydrogen graphitization method in closed reactors with heated iron catalyst and a cryogenic water trap (Kontuľ et al. 2017; Povinec et al. 2015, 2024). The mortar and plaster samples after cryo-breaking (sequential cooling with liquid nitrogen and heating to 80°C), were gently crushed and sieved. CO₂ was evolved from the samples in a glass vacuum line by hydrolysis using phosphoric acid. The CO₂ resulting after the first 7 seconds of hydrolysis was collected in liquid nitrogen and after purification was used to prepare graphite targets as already described. The prepared graphite samples were then sent for AMS measurement to the Athens laboratory.

Debrecen Laboratory at the Institute for Nuclear Sciences

Standard procedures were used for pretreatment of analyzed charcoal samples. Samples were treated with the ABA method. The prepared samples together with CuO were then sealed under vacuum in quartz tubes. The samples were combusted at 900°C, and then the CO₂ produced after purification was introduced into a glass graphitization reactor, containing clean zinc as a reducing agent (Molnár et al. 2013a; Rinyu et al. 2013, 2015). AMS measurements were carried out using the MICADAS (Ionplus, Dietikon, Switzerland) spectrometer (Molnár et al. 2013b; Orsovszki and Rinyu 2015). The NIST HOX II SRM 4990–C standard reference material was used for calibration (Schneider et al. 1995). Fossil carbon dioxide was used as a background sample.

Prague Laboratory at the Institute of Nuclear Physics

Standard procedures were used for pretreatment of analyzed wood and charcoal samples. After visual checks of the samples for possible impurities, the samples were treated using the already described ABA method. The prepared samples were then sealed off under dynamic vacuum into quartz ampoules with the addition of CuO. The samples were then combusted at a temperature of 900 °C. The purified carbon dioxide was then introduced into a graphitization reactor, which contained clean zinc as a reducing agent (Molnár et al. 2013a; Rinyu et al. 2013, 2015). The prepared graphite samples were finally vacuum sealed in glass ampoules and sent for AMS measurements in the Debrecen laboratory (Molnár et al. 2013b; Orsovszki and Rinyu 2015).

Tucson Laboratory at the University of Arizona

Standard procedures were used for the pretreatment of analyzed charcoal samples using the standard ABA method (Lange et al. 2019). The pretreated charcoal was combusted in a quartz tube with CuO. The resulting CO₂ was then purified and graphitized using the standard zinc-iron graphitization method. The prepared graphite samples were measured by AMS system with 3 MV NEC Pelletron accelerator (Jull et al. 2006, 2008). The measured ¹⁴C/¹³C ratio was evaluated using SRM4990 (oxalic-I) and 4990C (oxalic-II) standard materials and corrected for δ¹³C measured by an isotope mass spectrometer.

Vienna Laboratory at the University of Vienna

The ¹⁴C-dated charcoal samples in the VERA laboratory were pretreated with the standard ABA method (Wild et al. 2008), but for the first step with HCl under a higher temperature of 80°C was used instead of the usual 60°C. The pretreated samples were then combusted at 900°C in evacuated/sealed ampoules in the presence of CuO and silver wire, and finally graphite targets were prepared. AMS analyses were carried out at the VERA laboratory using the Pelletron 3 MV tandem accelerator (Steier et al. 2004).

Zurich Laboratory at the ETH

Before more sophisticated analyses, bulk mortar samples were analyzed for ¹⁴C and ¹³C content. The mortar samples were then dried sieved and after grain size separation, the fraction of 63–45 μm was used for the dissolution in 3-sec sequence intervals, with subsequent CO₂ collection in liquid nitrogen. The released CO₂ was collected in four separate fractions, plus a bulk CO₂ sample, and finally graphite targets were prepared (Hajdas et al. 2020). AMS analysis was carried out using the MICADAS AMS system (Ruff et al. 2010). The analytical method for the radiocarbon dating of mortar at ETH Zurich was described in detail by Hajdas et al. (2020). The charcoal sample was treated with the ABA method (Hajdas 2008) and analyzed by AMS as a graphite sample.

Calculation of radiocarbon ages

The measured ¹⁴C activities and their uncertainties are expressed as BP ages (before present, i.e., before 1950) using the Stuiver and Polach (1977) convention with a Libby ¹⁴C half-life of 5568 years. The combined uncertainties (statistical and systematic) of ¹⁴C ages are given as one sigma, corresponding to approximately 68% probability. The ¹⁴C ages expressed as BP represent uncalibrated ¹⁴C dates (Stuiver and Reimer 1993). Due to past ¹⁴C activity fluctuations in the biosphere, these ages should be calibrated using the IntCal20 calibration curve obtained from precise ¹⁴C measurements in tree-ring samples of known age (Reimer et al. 2020). The uncertainties of calibrated ages associated with the radiocarbon calibration curve are obtained from the conventional ¹⁴C ages and their combined uncertainties, which are then converted into calibrated calendar dates with the OxCal computer program (Bronk Ramsey 2020). Calibrated ages are expressed as age intervals in AD. The corresponding uncertainties are usually given at 2σ, representing a probability of 95.4% of the given age interval. Although the ¹⁴C ages of dated wood, charcoal, mortar and plaster samples quoted as BP have very small uncertainties (about ±15 yr for the average ages at 1 sigma), due to the specific plateau character of the calibration curve from the 9th to the 10th century, the calibrated ages have wider age intervals (as discussed later).

Results

The results of the radiocarbon dating of wood and charcoal samples collected from the Church of St. Margaret of Antioch in Kopčany, obtained from the international consortium laboratories (Table 1), show that a reasonable agreement has been obtained between the participating laboratories for given types of samples. It is important to note that the results did not show any inconsistencies between the samples and the laboratories. They agreed within the uncertainties quoted, and the results from different laboratories for wood and charcoal samples (except C-7) passed the χ²-tests. Therefore, we calculated the combined ages for these samples and proceeded with calculations of calibrated ¹⁴C age ranges with the OxCal v4.4.4 calibration program (Bronk Ramsey 2020) using the IntCal20 calibration curve of Reimer et al. (2020). As the C-7 data did not pass the χ²-test, they were not used in further evaluations.

Table 1. Results of ¹⁴C dating of wood (W), charcoal (C), mortar (M), and plaster (P) samples from the Church of St. Margaret of Antioch in Kopčany

¹ Uncertainties at 1σ.

² Average age and results of χ²-test (T-test statistics with the critical value for 5% significance level are given in the parentheses).

³ Calibrated calendar dates calculated with the OxCal computer program (Bronk Ramsey 2020) using the IntCal20 calibration curve (Reimer et al. 2020). Probability ranges at 95.4% (2σ confidence levels).

⁴ University of Georgia, Athens (Georgia, USA).

⁵ The combined results of C-7 sample did not pass the χ² test, χ² T = 20.2 (3.8).

⁶ The combined results of the first two M-2 fractions did not pass the χ² test, χ² T = 18.6 (3.8).

⁷ The combined results of the first two M-3 fractions did not pass the χ² test, χ² T = 43.1 (3.8).

Radiocarbon dating of wood and charcoal samples

Three laboratories (Athens, Bratislava and Prague) provided ¹⁴C results for the wood sample that were in good agreement. They passed the χ²-test and the combined ¹⁴C age is 1197 ± 12 BP. The calibrated ¹⁴C age of the wood sample is 774–884 AD (95.4% confidence level) (Table 1, Figure 6A).

Figure 6. The OxCal ¹⁴C ages of the wood sample (A), combined charcoal samples (B) and combined wood and charcoal samples (C).

Six laboratories provided ¹⁴C results for eight charcoal samples from which data sets for C-7 sample did not pass the χ²-test. However, the rest of the results (C-1, C-3, C-4, C-5 and C-6) were in very good agreement (they passed the χ²-test), and, as can be seen in Figure 6B, we could obtain from the combined charcoal results the calibrated ¹⁴C age of 773–885 AD (95.4% confidence level). Samples C-2 and C-9 showed younger (1110 ± 15 BP) or older (1327 ± 15 BP) ages, respectively, and since they did not pass the χ²-test, when compared to the rest of the charcoal samples, they were excluded from the calculation of the combined calibrated ¹⁴C age of charcoal samples.

If we combine the ¹⁴C results of wood and charcoal samples together (they passed the combined χ²-test), we get the calibrated ¹⁴C age of 773–885 AD (95.4% confidence level) (Figure 6C), that is, exactly the same age as for the combined charcoal data set. This result supports the idea that the old wood effect in the combined charcoal samples is negligible. The relatively wide interval of the calibrated ages of wood and charcoal samples, as well as the calibrated ages of combined wood and charcoal samples, is a consequence of the specific character of the calibration curve in this time period (Figure 6).

Radiocarbon dating of mortar and plaster samples

Two mortar samples and one mortar plus plaster sample were analyzed by the Zurich and Bratislava/Athens laboratories, respectively (Table 1). The Zurich laboratory analyzed the mortar samples using the sequential method with absorption of released CO₂ per partes every 3 seconds (Table 2). Usually, the 1st and the 2nd fractions give the best results as the 3rd and 4th fractions could be more influenced by fossil calcite, as documented in Figure 7. For the M-2 sample the combined results for the 1st (1091 ± 21 BP) and the 2nd (1219 ± 21 BP) fractions did not pass the χ²-test, therefore, we used only the first fraction to calculate the calibrated ¹⁴C age (Hajdas et al. 2020). Similarly, in the case of the M-3 sample, we cannot combine the first two fractions, as they did not pass the χ²-test (Table 2).

Table 2. Results of the sequential ¹⁴C analysis of mortar (M) samples performed by ETH Zurich

¹ Uncertainties at 1σ.

² The combined results of the first two M-2 fractions did not pass the χ² test, χ² T = 18.6 (3.8).

³ The combined results of the first two M-3 fractions did not pass the χ² test, χ² T = 43.1 (3.8).

Figure 7. Radiocarbon ages of mortar samples measured by the Zurich group.

A dating of the M-1 sample carried out by the Bratislava group (CO₂ from the first 7 seconds of the sample dissolution) gave a result of 1221 ± 20 BP, in agreement with wood and charcoal samples. The calibrated ¹⁴C age of the M-1 sample is 706–883 AD (95.4% confidence level) and fits again very well with the wood and charcoal samples (Table 1).

The internal plaster sample was taken from the second surface plaster layer and, therefore, it was expected that it would not be an original plaster but of secondary origin after some reconstruction work carried out in the past. As we can see from Table 1, the conventional age of the P-1 sample is 985 ± 20 BP and its calibrated ¹⁴C age is 997–1154 AD (95.4% confidence level), much younger compared to the ages of wood and charcoal. Unfortunately, we were not allowed to carry out more sampling in the interior of the Church, not to damage recently made new plaster cover.

Discussion

Radiocarbon age of the church

Figure 8 shows the calibrated age ranges of wood, charcoal, mortar, and plaster samples. It can be seen that the age ranges are generally well overlapping. The best data set was obtained for one wood and five charcoal samples (all data passed the χ²-test), giving the combined calibrated age range of 773–885 AD (95.4% confidence level). The older charcoal sample C-9, 655–774 (95.4% confidence level) (Table 1) may represent the effect of old wood, which was used during construction work. If an old tree was used to make a fire for preparation of lime, and a charcoal sample used to date was from the central part of the tree, the corresponding age could be too old compared to the true age of the dated object. This age uncertainty should always be in mind when dating objects on the basis of wood/charcoal dating (Povinec et al. 2021).

Figure 8. Comparison of OxCal ¹⁴C ages for charcoal (including combined results), wood, mortar, and plaster samples (774–884 AD (95.4% confidence level) for the wood sample, 773–885 AD (95.4% confidence level) for combined charcoal samples, and 773–885 AD (95.4% confidence level) for the combined wood and charcoal samples).

Fortunately, we were able to date a piece of wood of a diameter of 2 cm only, which was used during the construction of the Church. In the case when we take only the wood sample as the best indicator of the age of the Church, representing a single piece of wood with five annual rings (compared with charcoal pieces where we cannot determine annual rings), the calibrated ¹⁴C age is 774–884 AD (95.4% confidence level).

Unfortunately, due to sampling restrictions, the sampling points of the charcoal samples found in the mortar could only be approximately paired with the mortar samples. The C-3 charcoal sample found in the southern wall, 70 cm above the ground, could be approximately paired with the M-1 mortar sample taken at 80 cm. Although the distance between the sampling points was 150 cm, their ages (1167 ± 16 BP and 1221 ± 20 BP, respectively) are within statistical uncertainties. The C-7 charcoal sample found on the eastern wall, 150 cm above the ground, could be compared with the M-2 mortar sample found 100 cm above the ground. The resulting ages of 1184 ± 14 BP and 1091 ± 21 BP, respectively, are again within statistical uncertainties. The C-6 charcoal sample found in the northern part of the apse wall, 110 cm above the ground, could be compared with the M-3 mortar sample taken in the northern part of the apse wall at 80 cm above the ground. In this case, the resulting ages of 1203 ± 15 BP and 837 ± 21 BP indicate that the M-3 sample could probably be influenced by reparation work carried out in 1926 when the northern wall of the Church was repaired with cement-lime mortar and all walls were coated with a new layer of plaster. After two decades, the plaster fell again, leaving the wall open with cracks in the mortar for nearly 80 years, which were favourable for recrystallization. In the presence of rainwater, the mortar binder CaCO₃ can dissolve, react with fresh atmospheric CO₂ and deposit. Recrystallized CaCO₃ then has ¹⁴C age younger than the time of construction (Daugbjerg et al. 2021). This could explain why the mortar sample collected on the north wall (M-3) showed a much younger age compared to the samples M-1 and M-2.

An interesting case is the C-9 charcoal sample found inside the Church, buried inside the mortar, just below the plaster layer representing the sample P-1. In this case, the resulting ages of 1327 ± 15 BP and 985 ± 20 BP, respectively, indicate that the P-1 sample should represent a newer plaster layer introduced during the Church repair.

The ¹⁴C results reported in this work confirmed the usefulness of the mortar dating, especially if such data could be supported by wood and/or charcoal radiocarbon data. As expected, the dating results of mortar samples presented a wider range of conventional ages than wood and charcoal samples, documenting that radiocarbon dating of mortar is still a challenging process. Recently, new developments have been reported (Hajdas et al. 2020; Urbanová et al. 2020), which would further improve the radiocarbon dating of mortar.

Bayesian model

Although, there has been good agreement in the data sets and their evaluation using the OxCal calibration method, indicating that the Church of St. Margaret of Antioch in Kopčany was built in the ninth century, we also carried out the Bayesian modeling to get more robust results which would confirm these estimations. Bayesian modeling, based on the MCMC (Markov Chain Monte Carlo), plays a crucial role in data analysis because of direct calculation of the posterior distribution in complex models. It has recently been frequently used also in archaeology, especially when dealing with large data sets of different origin (Bayliss 2007, 2009; Bronk Ramsey 2009, 2017).

Figure 9 shows the results of Bayesian modeling of wood, charcoal and mortar samples arranged in two Phases. Phase 1 included cutting of wood and lime burning, which could be represented by wood and charcoal samples. The Phase 2 included construction works and the mortar setting, which could be represented by mortar samples. The plaster sample represents the second layer plaster and therefore it has not been included in the calculations.

Figure 9. The output plots from the Bayesian model for dating wood, charcoal and mortar samples (the old charcoal outlier, C-9, with ¹⁴C calibrated age range of 655–774 AD (95.4% confidence level) was used as the terminus post quem). The Boundary Transition 1/2, interpreted as the possible construction period of the Church, is estimated to be in the range of 788–884 AD (95.4% confidence level).

There are three important time marks in Figure 9 represented by the Boundary Start, the Boundary End and the Boundary Transition 1/2. The Boundary Start (the beginning of Phase 1) is represented by the wood and charcoal samples (774–877 AD); the Boundary End (the end of the Phase 2) is represented by the mortar samples (883–1161 AD); and the transition period between these two phases (the Boundary Transition 1/2) is interpreted as the possible construction period of the Church (Bayliss 2015; Bronk Ramsey 2017). The old charcoal outlier (C-9) has been used as a terminus post quem (655–774 AD), and the young charcoal outlier (C-2) as a terminus ante quem (892–992 AD) constraints for Phase 1 in the model.

The resulting Boundary Transition 1/2 is estimated to be in the range of 788–884 AD (95.4% confidence level), which fits very well with the calibrated age of 774–884 AD (95.4% confidence level) obtained for the wood sample, but also with the combined wood and charcoal age of 773–885 AD (95.4% confidence level). The Bayesian model also includes calculation of the old wood effect which may affect the ages of the charcoal samples (the wood sample was a type of branch with diameter of 2 cm only). The results presented in Figure 9 were modeled for possible inbuilt age with the help of Exponential Outlier model (Dee et al. 2014), which assumes an age of 1–100 yr. The absence of the old wood effect in the Bayesian modeling results thus confirms the results obtained in OxCal calibration that the old wood effect in the combined charcoal samples is negligible.

We may therefore conclude that the ¹⁴C age range of the construction of the Church of St. Margaret of Antioch in Kopčany obtained by Bayesian modeling (788–884 AD, 95.4% confidence level) is very close to the OxCal age of 774–884 AD (95.4% confidence level) obtained for the wood sample, as well as for the combined charcoal and wood plus charcoal combined results.

The old charcoal outlier (C-9) that has been used as the terminus post quem in the Bayesian modeling (Figure 9) has a wide range of 655–774 AD (95.4% confidence level), influenced by the plateau character of the IntCal20 curve (Figure 6). This has been shifting the age of the Church to dates even before the end of the 7th century, which has also been influencing the Bayesian modeling. According to written historical documents (Anonymous 870), Salzburg Archbishop Adalram consecrated in 828 AD a church in Nitrava Castle, which was built by Prince Pribina, a ruler of the Principality of Nitrava. It is natural to expect that a church in the Nitrava Principality capital would be built earlier than a church in the countryside. We may propose therefore to use this documented date as the terminus post quem and carry out the Bayesian modeling (Figure 10). In this case, the Boundary Start (the beginning of Phase 1, represented by the wood and charcoal samples, 774–877 AD), will change to 825–871 AD, and the Boundary Transition 1/2 will be in the range of 837–884 AD (95.4% confidence level), which, as expected, much better reflects the historical expectations. The Bayesian modeling thus improves the interpretation of ages which could be associated with known historical events.

Figure 10. The output plots from the Bayesian model for dating wood, charcoal and mortar samples using the consecration date of the church in Nitrava (828) AD as the terminus post quem. The Boundary Transition 1/2, interpreted as the possible construction period of the Church, is estimated to be in the range of 837–884 AD (95.4% confidence level).

Radiocarbon calibration curve effect

As can be seen in Table 1, the AMS technology enables to date wood and charcoal samples with high precision, where ¹⁴C conventional ages around 1200 BP, with 1σ uncertainties of single measurements around ±20 yr (with the weighted mean averages only around ±15 yr) have been obtained. The specific character of the ¹⁴C calibration curve during the 9th century (Figure 6) has a negative impact on the precision of calibrated ages, as the range of AD calibrated ages due to the ¹⁴C plateau is 60–80 years. This behaviour of the calibration curve with resulting uncertainties of calibrated ages prevents establishing a more precise date of the historical events that occurred during the 9th century. Fortunately, the new calibration curve IntCal20 improved the situation when compared with the old IntCal13; however, further improvements of the calibration curve are still needed, specifically in this time interval.

Historical comments

Recently, the same international radiocarbon consortium dated the Rotunda of St. George in Nitrianska Blatnica (Povinec et al. 2021). The calibrated ¹⁴C age of the Rotunda using the oldest calibrated age range obtained for a wood sample (measured with high precision) was 774–880 AD (95.4% confidence level). Combining the data set obtained for four wood samples (passing χ²-test), the calibrated age range of 773–887 AD (95.4% confidence level) was obtained. The Bayesian modeling carried out for the Rotunda (Boundary Start at 761-874 AD and Boundary End at 893-1039 AD) gave the Boundary Transition 1/2 range of 785–886 AD (95.4% confidence level), which is almost the same as for the Kopčany Church, 788–884 AD (95.4% confidence level) (both modeled for possible inbuilt age with the help of Exponential Outlier model, which assumes an age of 1–100 yr). If we use as the terminus post quem the consecration date of the Nitrava church (825 AD), the Boundary Start will change to 825–868 AD, the Boundary End to 893–1042 AD, and the Boundary Transition 1/2 of the Rotunda will be in the range of 837–881 AD (95.4% confidence level), which is in a very good agreement with the range of 837–884 AD (95.4% confidence level) determined for the Church in Kopčany.

We can see that the dating results of the Rotunda of St. George in Nitrianska Blatnica gave almost the same results as for the Church in Kopčany; therefore, we could expect that they were constructed during a similar time. From a historical perspective, it is interesting whether the Kopčany Church was built before or after the arrival of Constantine (St. Cyril) and St. Methodius to Great Moravia in 863, because they represent a significant milestone in the Christian history of Central Europe. The probability distribution from the radiocarbon calibration shows that the Church in Kopčany and the Rotunda in Nitrianska Blatnica were built in the period before 863 AD with a probability of 79% and 86%, respectively. The Bayesian modeling suggests that for the age ranges of 837–884 AD and 837–881 AD, the churches were built before 863 AD with a probability of 64% and 65%, respectively.

The Church of St. Margaret of Antioch in Kopčany together with the Rotunda in Nitrianska Blatnica are at present the only known churches built in Great Moravia that have been preserved in the original sacral structure. Thanks to the degree of preservation, they represent unique monuments and proofs of Christian cultural traditions forming the Central European region. In the framework of the wider Mikulčice residential agglomeration, the Church in Kopčany is the only standing Great Moravian building, which provides an important database of observations for the interpretation and reconstruction of Great Moravian sacred buildings. Its relatively precise radiocarbon dating can be considered as an important contribution to systematic investigations of the sacral architecture of the Great Moravia. Unfortunately, the large uncertainty in radiocarbon calibrated ages influences research on the history of Great Moravia, which plays an important role in Czech and Slovak history. Therefore, a more precise ¹⁴C calibration curve for the 9th century is under development, which would improve the calibrated age data.

Conclusions

The international consortium of radiocarbon laboratories (Athens and Tucson (USA), Debrecen (Hungary), Prague (Czech Republic), Vienna (Austria), Zürich (Switzerland) and Bratislava) was established in 2021 with the aim of dating the origin of the Church of St. Margaret of Antioch in Kopčany, and to contribute to better understanding of the origin of Christianity in Great Moravia. In total, 13 samples of wood, charcoal, mortar, and plaster were analyzed. The results of ¹⁴C measurements of the international consortium showed consistency between laboratories when analyzing samples of different origin. The obtained radiocarbon data sets were in good agreement, resulting in well-established radiocarbon calibrated ages.

In the case where we shall take only the wood sample dating results as the best indicator of the age of the Church (representing a single piece of wood with five annual rings only), the calibrated ¹⁴C age is 774–884 AD (95.4% confidence level). Five combined charcoal results (they passed the χ²-test) gave the calibrated ¹⁴C age of 773–885 AD (95.4% confidence level). When we combine radiocarbon results for the wood sample and charcoal samples (all data passed χ²-test), the combined calibrated age range is 773–885 AD (95.4% confidence level). We may conclude, therefore, that on the basis of the radiocarbon dating of the wood sample and the OxCal calibration, the Church of St. Margaret of Antioch in Kopčany was constructed in the 8th–9th century (774–884 AD (95.4% confidence level). The results of the Bayesian modeling give an age range of 788–884 AD (95.4% probability), which is very close to the calibrated age of the wood sample (as well as of the charcoal samples). If we take as the terminus post quem the date of the documented consecration of the church in Nitrava (828 AD), the Bayesian modeling suggests the age range of 837–884 AD (95.4% confidence level), which even better reflects the historical expectations.

Although the experimental uncertainties associated with ¹⁴C determinations were reasonably low (±20 yr for single measurements), the wide resulting calibrated age range has been heavily influenced by the specific shape of the radiocarbon calibration curve.

The radiocarbon dating results of the St. Margaret of Antioch Church in Kopčany, together with the St. George’s Rotunda in Nitrianska Blatnica (Povinec et al. 2021), confirmed that they probably represent the oldest standing purpose-built Christian churches in Central Europe east of Salzburg.