The medieval dietary transition in the ecclesiastical centre of Stara ´ Boleslav (9th-15th centuries AD, Czechia)

The medieval dietary transition in the ecclesiastical centre of Star´ a Boleslav (9th-15th centuries AD, Czechia) Sylva Drtikolov´ a Kaupov´ a a,b,* , Petra Str´ ansk´ a a,b , Petr Velemínský a,b , Ivana Boh´ aˇ cov´ a b a Department of Anthropology, National Museum, V´ aclavsk´ e n´ amˇ estí 68 Prague 1 11579, Czechia b Institute of Archaeology CAS, Letensk´ a 123/4 118 01 Prague, Czechia ARTICLE INFO Keywords: Stable isotopes Middle Ages Millet Freshwater fish Central Europe ABSTRACT In order to investigate dietary changes during the Middle Ages, carbon and nitrogen isotopes were measured in 97 humans and 14 animals from the medieval centre of Star´ a Boleslav, together with a comparative dataset of 28 humans from Prague – Loret´ ansk´ e Square. The Star´ a Boleslav dataset was divided into three contexts based on grave location (abandoned unknown Early Medieval church, and collegiate chapter basilica) and chronology (preand post-1150 CE phases of the basilica burial ground). The average human ẟ 13 C value was −19.2 ±0.4 ‰ and the average ẟ 15 N value was 10.3 ±1.2 ‰. There were statistically significant differences in ẟ 15 N values between all contexts with the exception of the abandoned church and the Prague – Loret´ ansk´ e Square cemeteries, which showed similarly low ẟ 15 N values. The post-1150 CE phase of the basilica burial ground showed the highest ẟ 15 N values, while the pre-1150 phase of the basilica burial ground occupied the intermediate position. The results show the almost complete absence of millet in the diet of the studied dataset, suggesting that a dietary dichotomy in millet consumption was present in the 11th century population, with the inhabitants of the centres consuming less millet than observed in previously published rural datasets. The increase in the consumption of animal products and/or fish was pronounced in the most recent phase of the dataset, after 1150 CE. Some evidence of this pattern was present even earlier, but only in the environment closely linked to the ecclesiastical centre, i.e. the basilica. 1. Introduction As a complex bio-socio-cultural activity, human dietary behaviour reflects not only the nutritional needs of individuals, but also a range of cultural, socio-economic and religious beliefs, rules and taboos. (Adamson 2004). In the Central European context, the Middle Ages are earliest historical or prehistoric context to provide numerous skeletal series large enough to allow the study of the impact of these factors on human diet. Previous research, using the well-established technique of analysing stable carbon and nitrogen isotopes in bone collagen, has provided a detailed description of the Early Medieval diet (9th-11th centuries AD) and described profound socio-economic differences in access to food resources (Kaupov´ a et al. 2018; 2019; Koˇ stov´ a et al. 2022). These concerned not only animal products, which is a common finding in the context of medieval populations (e.g. Colleter et al. 2019; Hakenbeck et al. 2010; P´ erez-Ramallo et al. 2022; Reitsema and Vercellotti 2012; Yoder 2012), but also the composition of the plant part of the diet, namely the proportion of millet, which was mainly consumed by members of lower socio-economic groups. Further research began to cover the High Medieval period (i.e. the 13th to 14th centuries AD; Drtikolov´ a Kaupov´ a et al. 2023; Orna et al. 2024), describing another phenomenon observed throughout Europe (Aguraiuja-L¨ atti and L˜ ougas 2019; Curtis-Summers et al. 2020; Mion et al. 2019; Müldner and Richards 2007) the increase in the consumption of animal products and/ or fish during the Middle Ages. However, further isotopic exploration of the numerous faunal datasets (Kovaˇ cikov´ a et al. 2020) suggested that the underlying cause of the observed human isotopic patterns may well be of a more complex nature, related to the profound agro-economic transformation of the Czech Lands during the Middle Ages (Kl´ apˇ stˇ e 2005). Thus a substantial white space on the isotopic map of medieval Czechia remains to be filled in order to understand the ’medieval dietary transition’ in this area. As a next step in filling in this map, this paper * Corresponding author at: Department of Anthropology, National Museum, V´ aclavsk´ e n´ amˇ estí 68 Praha 1 11000, Czechia. E-mail addresses: [email protected] (S. Drtikolov´ a Kaupov´ a), [email protected] (P. Str´ ansk´ a), [email protected] (P. Velemínský), [email protected] (I. Boh´ aˇ cov´ a). Contents lists available at ScienceDirect Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep https://doi.org/10.1016/j.jasrep.2025.105299 Received 14 January 2025; Received in revised form 13 June 2025; Accepted 4 July 2025 Journal of Archaeological Science: Reports 66 (2025) 105299 Available online 11 July 2025 2352-409X/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). presents the results of a dietary reconstruction of the 9th to 15th century population of Star´ a Boleslav, offering a unique opportunity to observe the whole process of the transition at one archaeological site. Star´ a Boleslav was founded around 900 CE as one of the strongholds on which the ruling Pˇ remyslid dynasty relied to build the early Czech state. Before the middle of the 11th century (1039–46), the Bohemian prince Bˇ retislav I founded here a magnificent basilica and an important religious institution, the collegiate chapter. The site was probably chosen in connection with the cult of Prince Wenceslas, who was murdered here in 935AD and later became the patron saint of the Bohemian nation and the most important Czech saint. Star´ a Boleslav underwent a significant transformation in connection with the founding of the chapter: from being a dynastic centre with a small church and settlement, it eventually became an important ecclesiastical locality completely different in appearance (Boh´ aˇ cov´ a 2023). From an archaeological point of view, it is one of the best preserved Early Medieval sites in the Czechia; the building activity that took place here in more recent periods affected the Early Medieval cultural layer only minimally (Boh´ aˇ cov´ a 2003; 2006; 2011; 2023). Undisturbed archaeological contexts from more than five centuries offer a unique opportunity to refine the timing of the medieval dietary transition and to study changes in dietary behaviour in relation to the changing function of the site. 1.1. Stable isotopic analysis and its application in reconstructing medieval diets Carbon and nitrogen isotopic analysis of bone collagen has become an established part of bioarchaeological studies over the last two decades, allowing the reconstruction of human dietary behaviour at the individual level. Carbon isotopic values (δ 13 C) differ between terrestrial and marine environments due to the different carbon sources. In a terrestrial environment, δ 13 C values differ between plants of the C3 and C4 photosynthetic pathways due to the different discrimination of 13 C during CO 2 fixation; these differences are naturally reflected in the tissues of consumers (Ambrose and Norr 1993; DeNiro and Epstein 1978; Lee-Thorp 2008). The latter phenomenon takes on great significance in the context of medieval Central Europe, where the decline of millet cultivation was suggested by earlier archaeobotanical findings (Hoffmann 2009). Thanks to the fact that both carbon and especially nitrogen (δ 15 N) isotopic values increase significantly along the food chain, the dietary proportion of animal products can be estimated (DeNiro and Epstein 1981; Lee-Thorp 2008; Minagawa and Wada 1984). While the bone collagen of herbivores has δ 13 C values approximately 5 ‰ higher than their forage (Ambrose et al. 1997), at higher levels of the food chain the difference in collagen δ 13 C values between prey and consumers is about 1 ‰ (Bocherens and Drucker 2003; Lee-Thorp 2008). For δ 15 N values, trophic level steps are much higher, with controlled feeding experiments with different animal species and direct observation for humans giving results in the range of 3–6 ‰ (Fernandes et al. 2012; Hedges and Reynard 2007; O’Connell et al. 2012). The dietary proportion of animal products is an important indicator of dietary quality, and has repeatedly been found to be linked to the social standing of the individual from prehistory (Hakenbeck et al. 2010; Knipper et al. 2015; Reitsema and Vercellotti 2012; Yoder 2012) to today (Adams 2018; Leroy and Praet 2015). Another food group that may have been particularly important in medieval Central Europe is freshwater fish, as the traditional explanation for the increase in nitrogen isotopic values observed throughout Europe during the Middle Ages is the increased consumption of fish due to stricter adherence to Christian doctrine (Aguraiuja-L¨ atti and L˜ ougas 2019; Barrett and Richards 2004; Curtis-Summers et al. 2020; Mion et al. 2019; Müldner and Richards 2007). In general, freshwater organisms have high δ 15 N values and low δ 13 C values, but the isotopic values of freshwater organisms are highly variable, making the estimation of the role of fish in the human diet a challenging task (Dufour et al. 1999; Katzenberg and Weber 1999). 1.2. Material The analysis of the sources so far shows that the burials in Star´ a Boleslav are mostly concentrated in two areas with differences in burial rite (Fig. 1). One is around the two surviving sacred buildings, St Wenceslas’ Basilica and the adjacent Church of St Clement. The other is in the immediate vicinity of an abandoned and now totally destroyed single-nave church (perhaps dedicated to Our Lady and St George; Boh´ aˇ cov´ a 2023). Although the nature of the excavation, limited to planned utility trenches, did not allow for extensive exploration of the burial grounds, the total number of recorded grave pits — mostly containing remains in their primary position — currently amounts to as many as 413 across both areas (Boh´ aˇ cov´ a and Hruˇ skov´ a, 2025). A single-nave Romanesque church of unknown dedication was located approximately 50 m east of the basilica’s eastern end. Its discovery was entirely unexpected, as the presence of another early medieval structure in the immediate vicinity of the basilica had not been anticipated. The burial rite observed in the vicinity of the abandoned Early Medieval church, as well as the results of radiocarbon dating, point to the presence of Early Medieval graves exclusively. The graves, originally most likely belonging to a row cemetery, appear in up to fivefold superpositions and are generally situated within approximately 12 m of the church. The results presented here represent the isotopic values of individuals from dated graves and graves stratigraphically linked to them (N =40). The chronological range of the dated burials lies between the 9th century and 1166 CE (Boh´ aˇ cov´ a and Hruˇ skov´ a, 2025). The church was likely demolished already during the Middle Ages, with a significant portion of its masonry having been looted by that time. Radiocarbon analyses of selected skeletal remains at St Wenceslas’ Basilica indicate the presence of some graves that predate the basilica itself. These early burial activities may be related to the historically attested Church of St Cosmas and Damian, in front of whose entrance, according to legend, Prince Wenceslas died in 935. This church is believed to have stood in the eastern part of the basilica. However, as these comprise only a few graves, they were not assigned to a separate subgroup in our dataset. A marked increase in burial activity clearly follows the foundation of the basilica after the mid-11th century. It should be noted that in the immediate vicinity of the basilica and the Church of St Clement, graves were found in up to tenfold superpositions and without radiocarbon dating of the entire skeletal assemblage it is difficult to determine to which burial phase most of the graves belong. The skeletal assemblage examined was therefore concentrated in the vicinity of St Clement’s Church and around the concurrently built linear walls directly adjoining the basilica.(Fig. 1), where graves predating the construction of this church (N =25) can be identified with certainty, or at least with the highest probability (given their position in the stratigraphy). At the time of sampling, the construction of St Clement’s was tentatively dated to between 1100 and 1150 CE, with the mid-12th century as the upper limit, but subsequent radiocarbon dating results indicated it may also fall into the period shortly after the mid-12th century. For the sake of simplicity, however, this period will be referred to as ’pre-1150 ′ . For the skeletal material from the period after the construction of St. Clement’s (henceforth ’post-1150 ′ , N =32), in addition to radiocarbon dating a change in burial rite is also an indicator of grave age. Its transformations (the beginning of nailed coffins, other placement of the hands than along the torso) are dated to approximately the end of the Early or beginning of the High Middle Ages (Boh´ aˇ cov´ a and Hruˇ skov´ a, 2025). The end of the later burial phase can be assigned to the first half of the 15th century, in the period of the Hussite wars (Boh´ aˇ cov´ a and Koˇ stov´ a, 2024). Part of the skeletal assemblage from the burial ground at Loret´ ansk´ e Square in Prague (N =28) was used as a comparative dataset. The use of S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 2 this material, excavated in the 1930 s, is possible today thanks to the recent processing of primary archival documentation (Boh´ aˇ cov´ a and Blaˇ zkov´ a 2011). Only material classified on the basis of grave goods and stratigraphy into the Young Hillfort Period (2nd half of the 11th century −1st half of the 12th century) was used. The examined dataset comes from a suburb of an elite centre −Prague Castle. Beside this, several published datasets from Czech Early through High Medieval contexts were used for further comparison (Drtikolov´ a Kaupov´ a et al. 2023; Kaupov´ a et al. 2018; 2019; Koˇ stov´ a et al. 2022). It should be noted here that the size and representativeness of the sample was limited by several factors, including 1) the nature of the rescue excavations, which took place in current streets and only along the lines of the utilities being built, which naturally had a direct impact on the incompleteness of the skeletal remains examined; 2) the aforementioned uncertainties in the dating of a considerable part of the excavated skeletal remains; 3) in the case of the comparative dataset from Prague – Loret´ ansk´ e Square, research practices in the 1930 s resulted in the preservation of only certain parts of the skeletons, while Fig. 1. Geographical location of Star´ a Boleslav (red circle) and Prague (black circle) (A); plans of the abandoned church and adjacent cemetery (B) and St. Wenceslas Basilica, St. Clement Church, and adjacent cemetery (C); analysed pre-1150 CE graves shown in turquoise, analysed post-1150 CE graves in red, and non-analysed graves in beige. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 3 other parts were reburied; and 4) the age composition of the available skeletal material, with infants and children under the age of 4 excluded from the analysis due to the possible influence of breastfeeding and/or physiological stress on their isotopic values (Beaumont et al. 2015; Fuller et al. 2006). The human dataset was accompanied by 14 comparative samples of local fauna, including cattle (N =8), caprines (N =5) and horse (N =1). The faunal bones were excavated from two archaeological situations at Star´ a Boleslav stonghold, the first (N =7) dated to before 1050 CE, the second (N =7) to after 1050 CE. 2. Methods In the case of the skeletal material from the basilica burial ground, the assessment of demographic parameters (sex and age at death) was recently carried out by one of the authors (PS) using the standard methodology (Bruzek 2002; Lovejoy et al. 1985; Meindl and Lovejoy 1985; Murail et al. 2005; Saunders et al. 1993; Schmitt 2005; 2008; Szilvassy and Kritscher 1990; Ubelaker 1989; Walker 2008) In the case of skeletal material from the abandoned church burial ground and Prague – Loret´ ansk´ e Square, the information was taken from previous reports (Boh´ aˇ cov´ a and Blaˇ zkov´ a 2011; Stloukal 1998). Approximately 250 mg of cortical bone was sampled for stable carbon and nitrogen isotope analysis. Samples were preferentially taken from ribs, but due to the factors outlined above, other bones such as long bones, vertebrae or even skull fragments had to be sampled in a significant proportion of individuals. We avoided sampling bone parts with the lowest turnover rates, i.e., areas of the cranial base, where a proportion of collagen may retain isotopic signals from early childhood and thus be influenced by breastfeeding (Jørkov et al. 2009). Otherwise, we do not consider differences in bone turnover rates to cause significant bias in the observed dietary patterns, especially when taking into account the results of comparisons between adult and subadult individuals (see below). Bone collagen extraction proceeded according to the Longin (1971) method as modified by Bocherens (1992). The collagen samples were prepared at the Department of Anthropology, National Museum, Prague, CZ. Elemental Analysis – Isotope Ratio Mass Spectrometry (EA-IRMS) was performed at Iso-Analytical, Crewe, UK. Stable carbon and nitrogen isotopic compositions were calibrated relative to the VPDB and AIR scales using IAEA-CH-6 and IAEA-N-1 inter-laboratory comparison standards. Measurement uncertainty was monitored using in-house standards: IA-R068 (soy protein, δ 13 C V-PDB =–25.22 ‰, δ 15 N AIR =0.99 ‰), IAR038 (L-alanine, δ 13 C V-PDB =–24.99 ‰, δ 15 N AIR =–0.65 ‰), IA-R069 (tuna protein, δ 13 C V-PDB =–18.88 ‰, δ 15 N AIR =11.60 ‰) and a mixture of IAEA-C7 (oxalic acid, δ 13 C V-PDB =–14.48 ‰) and IA-R046 (ammonium sulphate, δ 15 N AIR =22.04 ‰). Precision was determined to be ±0.14 ‰ for both δ 13 C and δ 15 N values based on repeated measurements of calibration standards, check standards, and sample replicates. Accuracy or systematic error was determined to be ±0.07 for δ 13 C and ±0.11 for δ 15 N values based on the difference between the observed and known δ values of the check standards and the long-term standard deviations of these check standards. The total analytical uncertainty as defined by Szpak et al. (2017) was estimated to be ±0.16 ‰ for δ 13 C values and ±0.18 ‰ for δ 15 N values. The Wilcoxon-Mann-Whitney test or t-test was used to explore the isotopic variation within each burial site with respect to sex and age at death. Where appropriate, the exact version of the test was applied due to small sample sizes. Next, Welch’s ANOVA with Games-Howell posthoc test was used to analyse the differences between burial sites, as the assumption of homogeneity of variances was not met. Finally, to assess the correlation between stable isotope values and radiocarbon dating intervals (expressed as the minimum and maximum calibrated dates for each sample), we employed a Monte Carlo simulation approach (Robert and Casella 1999). This method appropriately accounts for the uncertainty and non-uniform distribution inherent in calibrated radiocarbon dates (Bronk Ramsey, 2016), followed by the calculation of Spearman’s rank correlation between the simulated dates and the isotopic values. All statistical analyses were conducted using R version 3.5.0. 3. Results All the samples provided sufficient collagen yield for stable carbon and nitrogen isotopic analysis. None of the samples was discarded after evaluation of the quality criteria (DeNiro 1985; Van Klinken 1999). The complete isotopic data can be found on the IsoArch platform (Plomp et al. 2024; Salesse et al. 2018) and in the online supplementary material 1. For the animal dataset (N =14; Fig. 2), ẟ 13 C values ranged from –22.7 ‰ to −20.1 ‰ with a median of −21.0 ‰, while ẟ 15 N values ranged from 4.2 ‰ to 9.4 ‰ with a median of 5.7 ‰. Neither ẟ 13 C values nor ẟ 15 N values differed significantly between the pre-1050 CE and post1050 CE samples (Wilcoxon-Mann-Whitney test; W =29, p =0.620 for ẟ 13 C values while W =33 and p =0.318 for ẟ 15 N values). There was also no statistically significant difference between the Star´ a Boleslav faunal dataset and previously published faunal data from early medieval Prague and its surroundings (Kaupov´ a et al. 2019; Wilcoxon–Mann–Whitney test: W =33, p =0.318 for δ 13 C values; W =133, p =0.215 for δ 15 N values). In humans (N =125; Table 1, Figs. 2 and 3), ẟ 13 C values ranged from −20.2 ‰ to −17.8 ‰ (mean =-19.2 ±0.4 ‰; median =-19.2 ‰), and ẟ 15 N values ranged from 7.6 ‰ to 13.9 ‰ (mean =10.3 ±1.2 ‰; median =10.0 ‰). In the abandoned church dataset (N =40), ẟ 13 C values ranged from −20.2 ‰ to −17.8 ‰ (mean =-19.3 ±0.5 ‰; median =-19.2 ‰), and ẟ 15 N values ranged from 7.6 ‰ to 11.9 ‰ (mean =9.4 ±0.7 ‰; median =9.3 ‰). There was no statistically significant difference between adults and subadults, nor between males and females (Table 2). In the basilica burial ground, older graves predating 1150 CE (N = 25) had ẟ 13 C values ranging from −19.9 ‰ to −18.5 ‰ (mean =-19.2 ± 0.3 ‰; median =-19.3 ‰), and ẟ 15 N values ranging from 8.7 ‰ to 13.9 ‰ (mean =10.6 ±1.2 ‰; median =10.3 ‰). It was not possible to compare the isotopic values between adults and subadults, as there were only two subadults in the dataset. There was no statistically significant difference between males and females (Table 2). In the younger part of the basilica burial ground (N =32), ẟ 13 C values ranged from −19.6 ‰ to −18.6 ‰ (mean =-19.1 ±0.2 ‰; median =-19.1 ‰), and ẟ 15 N values ranged from 8.9 ‰ to 13.3 ‰ (mean = 11.5 ±0.9 ‰; median =11.7 ‰). There was no statistically significant difference between adults and subadults or between sexes in either ẟ 13 C values or ẟ 15 N values (Table 2). Finally, the comparative dataset from Prague – Loret´ ansk´ e Square (N =28) showed ẟ 13 C values between −19.9 ‰ to −18.3 ‰ (mean =-19.1 ±0.4 ‰; median =-19.0 ‰), and ẟ 15 N values from 8.5 ‰ to 11.9 ‰ (mean =9.8 ±0.7 ‰; median =9.8 ‰). There was no statistically significant difference in either ẟ 13 C values or ẟ 15 N values between age or sex categories (Table 2). As there was no statistically significant variation in ẟ 13 C values or ẟ 15 N values in relation to biological parameters (sex and age-at-death) in any of the samples studied, differences between individual cemeteries were explored using one-way Welch’s ANOVA followed by intergroup comparison with post-hoc Games-Howell test (Table 3). Carbon isotopic values were almost identical in all of the cemeteries examined, with no statistically significant differences – Welch’s ANOVA, F(3, 62.75) = 1.26, p =0.296. The only statistically significant difference was the difference in variances (Levene’s test; p =0.007), with the abandoned church showing the highest variance. Nitrogen isotopic values, on the other hand, showed considerable variation, with the result of Welch’s ANOVA, F(3, 60.82) =39.08, p <0.001, indicating a statistically significant difference. Pairwise comparison showed statistically significant S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 4 differences between all pairs, except for the abandoned church and Prague – Loret´ ansk´ e Square, which showed similar (and the lowest) ẟ 15 N values. At the other end of the spectrum, with the highest ẟ 15 N values, lay the dataset from the post-1150 CE phase of the basilica burial ground. The pre-1150 graves from the basilica burial ground showed intermediate ẟ 15 N values. In the case of ẟ 15 N values, the Levene test also showed statistically significant variances between cemeteries, with the older phase of the basilica burial ground showing the highest variation of all the samples (p =0.027). For the 21 samples (Fig. 4) with both stable isotope values and radiocarbon dates (Boh´ aˇ cov´ a and Hruˇ skov´ a, 2025; Boh´ aˇ cov´ a and Koˇ stov´ a, 2024), the correlation between isotopic values and calibrated radiocarbon dates—accounting for chronological uncertainty via Monte Carlo simulation—was assessed using Spearman’s rank correlation. For δ13C values, the mean simulated Spearman’s rho was −0.51, with a 95 % confidence interval ranging from −0.70 to −0.31. The Monte Carlo p-value was 0.51, indicating no statistically significant correlation. For δ15N values, the estimated Spearman’s rho was 0.39 (95 % CI: 0.19–0.58), with a Monte Carlo p-value of 0.51, likewise suggesting a non-significant correlation. 4. Discussion The present data clearly show that the diet of the population samples studied was based primarily on C3 plants, with highly variable contributions from animal products and/or fish. Human-faunal carbon Fig. 2. Complete isotopic data from Star´ a Boleslav. Table 1 Basic statistics of human isotopic values. Context ẟ 13 C (‰)ẟ 15 N (‰) n Min-Max Median Mean ±SD Δ13 C h-fa Min-Max Median Mean ±SD Δ15 N h-fa The abandoned Church All 40 −20.2–17.8 −19.2 −19.3 ±0.5 1.8 7.6–11.9 9.3 9.4 ±0.7 3.4 Males 16 −20.2–18.8 −19.6 −19.5 ±0.4 1.6 7.6–10.3 9.2 9.1 ±0.6 3.2 Females 13 −20.0–18.0 −19.1 −19.2 ±0.5 1.9 8.8–11.9 9.5 9.7 ±0.8 3.7 Adults of ND sex b 5−19.9–17.8 −19.2 −19.1 ±0.8 2.0 8.9–10.3 9.3 9.5 ±0.6 3.5 Subadults 6 −19.4–18.5 −19.2 −19.1 ±0.3 2.0 8.9–10.5 9.4 9.5 ±0.6 3.5 Basilica, pre-1150 CE All 25 −19.9–18.5 −19.3 −19.2 ±0.3 1.9 8.7–12.6 10.3 10.6 ±1.2 4.6 Males 11 −19.9–18.6 −19.3 −19.3 ±0.3 2.0 8.9–11.7 10.6 10.5 ±0.7 4.5 Females 9 −19.6–18.5 −19.1 −19.1 ±0.3 1.8 8.7–12.6 10.2 10.3 ±1.3 4.3 Adults of ND sex b 3−19.7–19.3 −19.4 −19.4 ±0.1 1.7 9.6–12.6 11.0 11.0 ±0.9 5.0 Subadults 2 −19.7–19.0 −19.4 −19.4 ±0.5 1.7 10.8–12.1 11.4 11.4 ±0.9 5.4 Basilica, post-1150 CE All 32 −19.6–18.6 −19.1 −19.1 ±0.2 1.9 8.9–13.9 11.7 11.5 ±0.9 5.5 Males 15 −19.5–18.8 −19.1 −19.1 ±0.2 1.9 10.0–12.8 11.7 11.5 ±0.8 5.5 Females 6 −19.4–18.6 −19.1 −19.1 ±0.3 2.0 8.9–13.3 11.5 11.2 ±1.5 5.2 Adults of ND sex b 6−19.4–18.8 −19.2 −19.1 ±0.2 2.0 11.1–13.9 12.0 12.0 ±0.6 6.0 Subadults 5 −19.6–19.1 −19.4 −19.3 ±0.3 1.7 10.3–12.2 11.8 11.3 ±0.8 5.3 Prague – Loret´ ansk´ e Square All 28 −19.9–18.3 −19.0 −19.1 ±0.4 1.9 8.5–11.9 9.8 9.8 ±0.7 3.8 Males 5 −19.8–18.5 −19.0 −19.1 ±0.5 1.9 8.9–9.9 9.5 9.5 ±0.4 3.5 Females 9 −19.7–18.9 −19.3 −19.2 ±0.3 1.8 8.5–10.3 9.6 9.5 ±0.6 3.5 Adults of ND sex b 9−19.9—18.6 −19.0 −19.1 ±0.4 1.9 9.1–11.9 10.0 10.2 ±0.9 4.2 Subadults 5 −19.5–18.3 −18.9 −19.0 ±0.5 2.1 9.5–10.2 10.0 9.9 ±0.3 3.9 a human-faunal isotopic offset; as faunal data, the mean values of the main consumed domesticated species (cattle, caprines) were used. b adult individuals whose sex could not be estimated using osteological methods. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 5 isotopic offsets are below or just around 2 ‰ in the vast majority of individuals, which is considered the threshold for notable millet consumption. (Lightfoot et al. 2012). In fact, there are only two outliers in the whole dataset, both from the abandoned church burial ground (STB6/89 and STB20/91), for which substantial millet consumption can be suggested (Fig. 3). One of them (STB 20/91) is in fact the oldest of all the radiocarbon-dated graves (N =21), dating back to the 10th century CE (Fig. 4); unfortunately, the other was not radiocarbon dated. From Fig. 4 it is clear that there could be some minor input from millet in the oldest individuals, but there are almost no samples with carbon isotopic values above −19.0 ‰ dated after 1050 CE. As mentioned in the introduction, the abandonment of millet Fig. 3. Detailed view of human isotopic data from Star´ a Boleslav and Prague-Loret´ ansk´ e Square. Table 2 The results of t-tests and Wilcoxon-Mann-Whitney tests comparing δ 13 C and δ 15 N values between groups. Test statistics (t or Z), p-values, 95 % confidence intervals (CI), and mean or median differences are reported. Dashes (—) indicate insufficient data. Context Groups Compared Variable Test Type t / Z p 95 % CI Mean/Median Difference Abandoned church Female vs. Male δ 13 C Independent Samples t-test t =1.84, df =27 0.077 [-0.04, 0.64] 0.30 (mean) δ 15 N Independent Samples t-test t =1.81, df =27 0.081 [-0.07, 1.08] 0.51 (mean) Adult vs. Subadult δ 13 C Wilcoxon-Mann-Whitney Z =-1.27 0.214 [-0.62, 0.17] −0.27 (median) δ 15 N Wilcoxon-Mann-Whitney Z =-0.32 0.761 [-0.54, 0.40] −0.04 (median) Basilica, pre-1150 CE Female vs. Male δ 13 C Wilcoxon-Mann-Whitney Z =1.18 0.261 [-0.12, 0.49] 0.16 (median) δ 15 N Wilcoxon-Mann-Whitney Z =-0.84 0.423 [-1.45, 0.83] −0.42 (median) Adult vs. Subadult δ 13 C Wilcoxon-Mann-Whitney — — — — (insufficient data) δ 15 N Wilcoxon-Mann-Whitney — — — — (insufficient data) Basilica, post-1150 CE Adult vs. Subadult δ 13 C Wilcoxon-Mann-Whitney Z =1.82 0.070 [-0.03, 0.49] 0.20 (median) δ 15 N Wilcoxon-Mann-Whitney Z =0.23 0.835 [-0.56, 1.25] 0.07 (median) Female vs. Male δ 13 C Wilcoxon-Mann-Whitney Z =0.16 0.894 [-0.22, 0.33] 0.04 (median) δ 15 N Wilcoxon-Mann-Whitney Z =-0.47 0.677 [-1.50, 1.07] −0.23 (median) Prague – Loret´ ansk´ e Square Male vs. Female δ 13 C Wilcoxon-Mann-Whitney Z =0.60 0.582 [-0.50, 0.77] 0.11 (median) δ 15 N Wilcoxon-Mann-Whitney Z =-0.07 0.974 [-0.68, 0.82] −0.05 (median) Adult vs. Subadult δ 13 C Wilcoxon-Mann-Whitney Z =-0.66 0.529 [-0.62, 0.40] −0.10 (median) δ 15 N Wilcoxon-Mann-Whitney Z =-0.69 0.515 [-0.73, 0.44] −0.10 (median) Table 3 Games-Howell post hoc comparisons of δ 13 C and δ 15 N values between burial contexts. Reported values include the mean difference (M), 95 % confidence intervals (CI), adjusted p-values (Benjamini-Hochberg correction), significant results in bold. ẟ 13 Cẟ 15 N Comparison M 95 % CI pM 95 % CI p Abandoned Church vs Basilica, pre-1150 CE 0.077 [−0.20, 0.35] 0.880 1.15 [0.43, 1.87] <0.001 Abandoned Church vs Basilica, post-1150 CE 0.158 [−0.08, 0.40] 0.310 2.12 [1.59, 2.65] <0.001 Abandoned Church vs Prague – Loret´ ansk´ e Square 0.169 [−0.13, 0.46] 0.433 0.40 [−0.07, 0.86] 0.120 Basilica, pre-1150 CE vs Basilica, post-1150 CE 0.081 [−0.13, 0.29] 0.740 0.97 [0.19, 1.75] 0.010 Basilica, pre-1150 CE vs Prague – Loret´ ansk´ e Square 0.092 [−0.18, 0.37] 0.807 −0.75 [−1.50, −0.01] 0.046 Basilica, post-1150 CE vs Prague – Loret´ ansk´ e Square 0.011 [−0.23, 0.25] 0.999 −1.72 [−2.28, −1.16] <0.001 S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 6 cultivation during the Middle Ages has been documented both archaeobotanically and isotopically (Drtikolov´ a Kaupov´ a et al. 2023; Hoffmann 2009; Orna et al. 2024). However, the almost complete absence of the C4 plant isotope signal even in the oldest datasets from the abandoned church as well as in the comparative dataset from Prague –Loret´ ansk´ e Square is quite surprising, as we know from previous studies that millet was a common part of the diet, at least in the rural context, in the 11th century (Drtikolov´ a Kaupov´ a 2023; Kaupov´ a et al. 2018). Isotopic results from neighbouring Poland also date the decline of millet cultivation no earlier than the 12th century (Fetner 2025; Reitsema et al. 2017). Local archaeobotanical finds attest to the presence of millet in both Early and High Medieval contexts at the Star´ a Boleslav stronghold—including the acropolis, fortifications, and suburban zones. However, due to unfavourable soil conditions, the archaeobotanical assemblage from Star´ a Boleslav is extremely poor and does not permit an estimation of millet’s relative dietary significance (ˇ Culíkov´ a 2003). However, it is worth mentioning here the dichotomy in dietary habits that was observed in the Early Medieval (9th to 11th century) population. In the Early Medieval dataset, individuals from Prague Castle consumed significantly less millet than individuals buried in the Prague Castle hinterland and in rural areas (Kaupov´ a et al. 2019; Koˇ stov´ a et al. 2022). Since both Star´ a Boleslav and Prague – Loret´ ansk´ e Square are central locations, the current data suggest that the dichotomy in millet consumption described above persisted until the 11th century. The underlying causes of this pattern may be socio-economic, as millet was often considered a crop of lower socio-economic groups, mainly due to its low demands on soil quality (Adamson 2004; Weber and Fuller 2008). The absence of millet in the later time horizons represented by the individuals buried at St Wenceslas’ Basilica is consistent with existing indirect information on the diet of the medieval population, which points to the marginal role of millet in the 13th and 14th centuries (Hoffmann 2009). The decline of a popular Early Medieval crop may have been part of a wider change in agricultural technology and land use during the High Middle Ages (Kl´ apˇ stˇ e 2005). As millet is a thermophilic plant, climate change in the High Middle Ages may have been one of the causes, along with the expansion of agricultural areas to higher altitudes. Millet cultivation also requires a relatively large amount of human labour, which does not make it an ideal crop for feeding a growing population in central locations (Weber and Fuller 2008). Although a relatively high proportion of millet has been observed in some medieval urban Czech assemblages, this may be at least partly due Fig. 4. Scatter plot showing the relationship between calendar dates (mean of the 95% confidence interval) and isotopic values. Horizontal error bars represent the calibrated radiocarbon date ranges for each sample. The solid black regression line depicts the overall linear trend across all groups combined, with shaded confidence intervals illustrating the uncertainty around the fit. Points are colored according to archaeological context: Abandoned Church (red), Basilica pre-1150 CE (green), and Basilica post-1150 CE (orange). Dashed vertical lines indicate the proposed timing of the dietary transition. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 7 to the way it was processed: unlike other cereals, millet was cleaned (hulled) directly in the household, which led to its overrepresentation in kitchen waste (Winklerov´ a 2011). In terms of animal products and/or fish consumption, the Star´ a Boleslav dataset clearly met our expectations for medieval dietary transitions. The abandoned church, as well as the cemeteries of Prague’s Loret´ ansk´ e Square, showed relatively low δ 15 N values and human-faunal nitrogen isotopic offsets, comparable to those of 9th-11th century Prague Castle and its immediate hinterland. However, they are significantly higher than in most other Early Medieval samples, as well as in the 11thcentury rural dataset, suggesting the elite status of the inhabitants of Star´ a Boleslav (Drtikolov´ a Kaupov´ a 2023; Kaupov´ a et al. 2019; Koˇ stov´ a et al. 2022). The post-1150 phase of the basilica burial ground population sample, on the other hand, shows significantly higher δ 15 N values, comparable to the High Medieval populations of the 13th-14th centuries (Drtikolov´ a Kaupov´ a et al. 2023; Orna et al. 2024). The pre-1150 CE part of the basilica dataset shows intermediate δ 15 N values that differ significantly from both the post-1150 CE graves and those from the abandoned church. This is in good agreement with the dating of the datasets. Although the end of the older phase of the basilica burial ground coincides with the end of burial activity around the abandoned church (i.e. c.1150 CE), the basilica burial ground is −on average −more recent, with the core of use occurring after 1050 CE. At the same time, since the construction of St Clement’s Church, which divides the burial ground of the basilica into two phases, is dated to halfway through the 12th century or shortly thereafter, it could be said that the first evidence of the medieval dietary transition, whatever its causes, can be dated to before 1150 CE. The observation from Fig. 4 however shows that there are some individuals with δ 15 N values around 11 ‰ directly dated before 1150 CE buried at the basilica, while contemporaneous individuals from the abandoned church show δ 15 N values of 10 ‰ or less. All individuals dated after 1150 CE show significantly higher δ 15 N values. Unfortunately, there is a gap with no individuals dated closely post-1150 CE, the period of interest in the dataset (Boh´ aˇ cov´ a and Hruˇ skov´ a, 2025). The small size of the dataset currently precludes more in-depth analysis of radiocarbon and isotopic dates. As noted above, changing religious motivations with increasing pressure to conform to Christian rules are often suggested as the cause of the ’medieval dietary transition’ (Barrett and Richards 2004; Mion et al. 2019; Rutgers et al. 2009). In order to avoid eating terrestrial (i.e. warmblooded) animals, fish was an essential part of the diet during Advent and Lent, as well as during the regular fast on Friday, which was even called Fish Day (Kyselý et al. 2022). In the Czech context, however, fish were not consumed exclusively during fasting periods but were a regular part of the diet. Certain species—such as pike—were considered luxurious delicacies at noble courts (Kl´ apˇ stˇ e 2011; Sůvov´ a 2012). High medieval accounting records from the Augustinian monastery in Tˇ reboˇ n (1367–1369) and Karlˇ stejn Castle (1423–1434) show that fish were purchased or caught throughout the year, not only during fasting periods, and were often consumed fresh (“pisces recentes”; Kyselý et al. 2022).Historical sources indicate that the regular weekly supply was covered by local production.. In the Middle Ages, pond farming experienced a great boom on the territory of the Czechia (ˇ Sarapatka et al. 2014). However, it should be mentioned that herein lies the potential problem for isotopic diet reconstruction: because of the unpredictable but potential impact of anthropogenic influence on the aquatic web and fish feeding behavior in medieval ponds, we can hardly rely on published fish results from the Early Middle Ages or even older periods (e.g. Kaupov´ a et al. 2019). This topic requires further in-depth investigation which is however beyond the scope of this paper. At Star´ a Boleslav itself, the absence of systematic sieving and flotation has clearly led to the poor recovery of fish remains, which are limited to isolated specimens of pike (Esox lucius), chub (Leuciscus cephalus), larger cyprinids, and possibly a single anadromous salmonid (Salmo salar or Salmo trutta fario). As such, these finds cannot be considered a reliable reflection of the role of fish in the local diet (Kyselý 2003a). Unfortunately, these remains were not available for sampling. In the absence of fish remains from Star´ a Boleslav, the interpretation relies solely on previously published fish isotope data—bearing the aforementioned limitations—from early medieval Central European contexts in Czechia and Poland (N =27; mean δ 13 C =–25.0 ±1.9 ‰; mean δ 15 N =9.5 ±2.4 ‰; Kaupov´ a et al. 2018; 2019; Reitsema et al. 2017). Interpretation also draws on broadly accepted principles, according to which freshwater fish consumption is indicated by elevated δ 15 N values in combination with variable δ 13 C values. A human–faunal nitrogen isotopic offset exceeding 5 ‰ is typically considered evidence of substantial fish intake (Katzenberg and Weber 1999). This cut-off value was crossed by the individuals from the post-1150 CE phase of the basilica burial ground. However, another noteworthy limitation of the Star´ a Boleslav faunal dataset is that there is no statistically significant difference between the preand post-1050 CE samples, but the upper range of the dating of the younger dataset was not assessed with certainty, so it cannot be excluded that some variation in the faunal baseline during the duration of the site may have been missed. It should also be noted that the Star´ a Boleslav individuals do not exhibit a shift in δ 13 C values that would be expected based on the published fish data mentioned above. However, carbon isotopic values in freshwater fish are highly variable, and the “masquerading” of fish consumption through δ 13 C values—where fish intake is not clearly reflected in human carbon isotopic signatures—is a well-documented phenomenon (e.g. Lilak and Oras 2023). Due to the increased intensity of trade in preserved sea fish after 1000 CE (Barrett et al. 2004; Orton and Barrett 2016), the possibility of the consumption of sea fish must also be discussed. In Czech written sources, the first references to the presence of preserved salted sea fish − herring −appear in the 11th century, but become increasingly common from the 13th century (for review see e.g. Kyselý et al. 2022). However, there is no isotopic signature of marine fish consumption in the present dataset. Although the isotopic values of marine fish can vary considerably depending on their geographical origin, they always show more or less elevated carbon isotopic values together with high nitrogen isotopic values. For archaeological fish originating from the marine environments of western and northern Europe, δ 13 C values range from –16.0 ‰ to –11.3 ‰, while δ 15 N values vary between 11.2 ‰ and 16.8 ‰. The eastern Baltic region constitutes a partial exception, exhibiting markedly lower δ 13 C values (–18.3 ‰ to –14.9 ‰) and δ 15 N values (9.2 ‰ to 13.2 ‰). Nevertheless, even these values remain clearly distinguishable from those of both terrestrial and freshwater resources(Barrett et al. 2011; Hutchinson et al. 2015). This pattern was not reflected in the human isotopic data. This may be due to the fact that marine fish were probably not eaten regularly; rather, they were used to meet the increased demand during longer periods of fasting, namely Lent, when pond and/or river fishing was limited (Kyselý et al. 2022). An alternative source of animal protein—milk and dairy products—was mostly tolerated by medieval Catholic canonical rule (Friedberg 1879; Fry 1981) and is supported archaeologically by the predominance of cattle in Czech medieval assemblages (Kovaˇ cikov´ a et al. 2020; Sůvov´ a et al. 2018). However, milk and meat from terrestrial animals are indistinguishable in carbon and nitrogen isotopic records and therefore cannot account for the observed isotopic shift (Knobbe et al. 2006). Moreover, references to dairy products in relation to fasting in Czech written historical sources are scarce and predominantly indirect (Kl´ apˇ stˇ e, 2005), a notable contrast to the comparatively frequent and explicit mentions of fish. As an alternative explanation for the shift in human nitrogen isotopic values, the higher dependence of the inhabitants of the growing centres on pork should be considered. It is known from the context of medieval Prague that the isotopic values of pigs increased above the level of herbivores due to intensive breeding in the limited space of High Medieval central places (Kovaˇ cikov´ a et al. 2020). In the Czech region, S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 8 however, the results of environmental research and historical sources consistently point to the dominant role of cattle, even in central settlements and urban environments of the 13th and 14th centuries (Kovaˇ cikov´ a et al. 2020; Sůvov´ a et al. 2018). This is also reflected in the composition of the archaeozoological assemblage from Star´ a Boleslav (Kyselý 2003b). The continued dominance of cattle throughout the Czech Middle Ages makes this explanation less probable. As another possible explanation, changes in isotopic values at the base of the food chain should be considered. Traditionally, an increase in manuring rates has been viewed as part of a broader package of agricultural innovations during the High Middle Ages—alongside tool improvements, more diverse crop production, and the adoption of the three-field system (Kovaˇ cikov´ a et al. 2020). However, published nitrogen isotopic data from medieval plant macroremains—though the dataset remains limited (Dreslerov´ a et al. 2021; L´ atkov´ a et al. 2025)— suggest that high medieval grains exhibit δ 15 N values comparable to, or even lower than, those from early medieval central places. The latter have been interpreted as reflecting the influence of large quantities of organic matter generated by the population of these centres (Dreslerov´ a et al. 2021). Based on the current data, a change in agricultural technique does not appear to be a likely cause of the observed isotopic pattern. All in all, the consumption of freshwater fish seems to be the most likely explanation. In this light, it appears that the ’medieval dietary transition’ at Star´ a Boleslav may be somewhat accentuated by the presence of the collegiate chapter. Comparison of data from the abandoned church, Prague – Loret´ ansk´ e Square and the pre-1150 CE phase of the basilica burial ground (Fig. 3), as well as direct comparison of Table 4 Comparison of human isotopic data from the Central European medieval sites. Region Site Datation (AD) n Kontext δ 13 Cδ 15 NΔ 13 C humanfaunaa Δ 15 N humanfaunaa Ref. Czechia Star´ a Boleslav −The abandoned church 10th-1st half of the12th 40 centre −19.3 ± 0.5 9.4 ± 0.7 1.8 3.4 This study Czechia Star´ a Boleslav −Basilica, pre1150 CE 2nd half of the 11th1st half of the 12th 25 centre −19.2 ± 0.3 10.6 ± 1.2 1.9 4.6 This study Czechia Star´ a Boleslav −Basilica, post-1150 CE 2nd half of the 12th −1420 s 32 centre −19.1 ± 0.2 11.5 ± 0.9 1.9 5.5 This study Czechia Prague – Loret´ ansk´ e Square 2nd half of the 11th – 1st half of the 12th 28 centre −19.1 ± 0.4 9.8 ± 0.7 1.9 3.8 This study Czechia Prague Castle 9th-11th 19 centre −19.3 ± 0.6 10.4 ± 0.7 1.6 3.7 (Kaupov´ a et al. 2019) Czechia Prague −Triangl 10th 19 hinterland of the centre −18.6 ± 0.3 9.6 ± 0.7 2.2 3.0 (Kaupov´ a et al. 2019) Czechia Levý Hradec 2nd half of the 9th10th 25 centre −18.7 ± 0.5 9.2 ± 0.8 2.2 2.6 (Kaupov´ a et al.2019) Czechia Pˇ rezletice 10-11th 20 rural −18.7 ± 0.5 9.4 ± 0.5 2.1 2.8 (Koˇ stov´ a et al. 2022) Czechia Vrˇ sany 11th 36 rural −18.5 ± 0.5 9.6 ± 0.7 2.3 2.1 (Drtikolov´ a 2023) Czechia Kostice, Josefov 11th 31 Small centre −17.2 ± 0.5 9.5 ± 0.6 3.1 3.9 (Kaupov´ a et al. 2018) Czechia Plzeˇ n 13th-14th 51 urban −19.6 ± 0.3 11.9 ± 1.0 1.5 4.5 (Orna et al. 2024) Czechia Kutn´ a Hora 13th-16th 24 urban −19.3 ± 0.2 12.2 ± 0.5 1.7 4.3 (Drtikolov´ a Kaupov´ a et al. 2023) Czechia Oˇ skobrh 13th-14th 20 rural −19.4 ± 0.2 11.4 ± 1.2 1.6 3.5 (Drtikolov´ a Kaupov´ a et al. 2023) Polland Kaldus IV 11th 37 rural −18.5 ± 1.0 10.2 ± 0.8 2.7 3.3 (Reitsema et al. 2017) Polland Giecz 11th-12th 24 rural −18.9 ± 0.4 9.2 ± 0.5 2.3 2.3 (Reitsema et al. 2010) Polland Gruczno 1 12th 34 urban −19.8 ± 0.4 9.3 ± 0.6 1.4 2.4 (Reitsema et al. 2017) Polland Kaldus 1 12th-13th 30 rural −19.5 ± 0.4 10.2 ± 0.7 1.7 3.3 (Reitsema et al. 2017) Polland Kalisz-Zawodzie 12th-13th 27 elite −19.6 ± 0.5 10.1 ± 1,0 2.3 4.4 (Fetner 2025) Polland Gruczno 2 13th-14th 32 urban −19.9 ± 0.3 9.2 ± 0.8 1.3 2.3 (Reitsema et al. 2017) Polland Kalisz-Zawodzie 12th-13th 27 elite −19.6 ± 0.5 10.1 ± 1,0 2.3 4.4 (Fetner 2025) Germany Dalheim 11th 24 rural −20.0 ± 0.2 9.9 ± 1.0 1.4 4.2 (Olsen et al. 2018) Germany Berlin – Petriplatz 1100–1315 CE 35 urban −20.0 ± 0.3 11.4 ± 0.8 1.1 4.2 (Zechini et al.2021) Germany Haithabu 9th-11th 46 urban −20.0 ± 0.5 11.5 ± 1.4 x b x b (Grupe et al. 2017) Germany Schleswig Rathausmarkt – phase 1 1070–1140 CE 144 urban −19.5 ± 0.7 12.0 ± 1.3 x b x b (Grupe et al. 2017) Germany Schleswig Rathausmarkt – phase 2 1140–1210 CE 66 urban −19.7 ± 0.8 11.9 ± 1.1 x b x b (Grupe et al. 2017) Germany Schleswig St. Clements 1250–1350 CE 59 urban −19.6 ± 1.5 11.8 ± 1.5 x b x b (Grupe et al. 2017) a human-faunal isotopic offset; as faunal data, the mean values of the main consumed domesticated species (cattle, sheep/goat, pig) were used. b not given by the authors. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105299 9