The first isotopic insight into the Bronze Age dietary transition in the area of present-day Czechia

The first isotopic insight into the Bronze Age dietary transition in the area of present-day Czechia Sylva Drtikolov´ a Kaupov´ a a,* , Lenka Půtov´ a a,b , Michal Ern´ ee c , Petr Koˇ c´ ar c , Lenka Kovaˇ cikov´ a d , David Parma e , Milan Salaˇ s f , Jiˇ rí Unger c , Petr Velemínský a a Department of Anthropology, National Museum, V´ aclavsk´ e n´ amˇ estí 68, 110 00 Prague 1, Czech Republic b Institute for History of Medicine and Foreign Languages, First Faculty of Medicine, Charles University, U Nemocnice 4, 121 08 Prague, Czech Republic c Institute of Archaeology of the CAS, Letensk´ a 123/4, 118 00 Prague, Czech Republic d Laboratory of Archaeobotany and Palaeoecology, Faculty of Science, University of South Bohemia, Na Zlat´ e stoce 3, 370 05 ˇ Cesk´ e Budˇ ejovice, Czech Republic e Institute of Archaeological Heritage, Kaloudova 1321/30, 614 00 Brno, Czech Republic f Institute of Archaeology, Moravian Museum, Zelný trh 6, 659 37 Brno, Czech Republic ARTICLE INFO Keywords: Stable isotopes Carbon Nitrogen Millet Czech Republic Early Bronze Age Late Bronze Age ABSTRACT To describe changes in diet and/or agricultural practices over the course of the Bronze Age on the territory of present-day Czechia, carbon (δ 13 C) and nitrogen (δ 15 N) isotopic values were measured in bone or tooth collagen from 103 humans and 30 animals, together with 8 samples of charred cereal grains. The dataset was divided into four subgroups according to geographical and chronological contexts: Early (2200–1500 BCE) vs. Late (1300–1020 BCE) Bronze Age, and Bohemia vs. Moravia. Both carbon and nitrogen isotopic values showed statistically significant differences between the chronogeographic contexts. In accordance with archaeobotanical results, there was no significant input of millet in the Early Bronze Age sample (average human-faunal isotopic offsets of 0.5‰ for Bohemia and 1.3‰ for Moravia), whereas in the Late Bronze Age the significant input of millet is clearly documented (average human-faunal isotopic offsets of 3.9‰ for Bohemia and 5‰ for Moravia). The statistically significant difference between the two Late Bronze Age geographical units clearly shows that there was considerable variation in millet consumption across Czech territory. The relatively low δ 15 N values observed in Late Bronze Age contexts (average human-faunal isotopic offsets of 1.9 ‰ for Bohemia and 2.4 ‰ for Moravia) may indicate a certain restriction of high-quality diet, namely animal products, in individuals exempted from the regular burial treatment of cremation. 1. Introduction The Bronze Age (2200–800 BCE) has attracted considerable attention from the Czech scientific community for over a century. In the 20th century, research focused mainly on typological-chronological and cultural-historical development, which led to the establishment of the theory of the continuous emergence of complex chiefdom societies. The role of the new technology of bronze-casting was often emphasised as the driving force behind the development of craft and trade. Information on subsistence strategies was often used only illustratively or in discussions about the sedentary nature of societies. The information potential of human remains was often reduced to a basic demographic description. In the Central European context, this approach changed little or not at all between the 1970s and the beginning of the 21st century. (Jir´ aˇ n 2013; Pleiner and Rybov´ a 1978; Podborský 1993). Thanks to a considerable amount of new data from large-scale excavations and the application of new theoretical and methodological approaches, the simplistic picture of a linear development of chiefdoms towards complex civilisations is gradually being replaced by the idea of significant diversity linked at a higher level only by some widely shared cultural elements (Parma 2017). At the same time the strict vertical stratification of society has been questioned, based on the analysis of grave goods (Sosna 2009). With the changing paradigm there arises a need to rewrite in as much detail as possible the basic life and * Corresponding author. E-mail addresses: [email protected] (S. Drtikolov´ a Kaupov´ a), [email protected] (L. Půtov´ a), [email protected] (M. Ern´ ee), [email protected] (P. Koˇ c´ ar), [email protected] (D. Parma), [email protected] (M. Salaˇ s), [email protected] (J. Unger), [email protected] (P. Velemínský). Contents lists available at ScienceDirect Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep https://doi.org/10.1016/j.jasrep.2025.105335 Received 16 December 2024; Received in revised form 18 July 2025; Accepted 24 July 2025 Journal of Archaeological Science: Reports 66 (2025) 105335 Available online 31 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/ ). subsistence strategies for respective time frames and regions, and to update the recognized general trends of this Epoque, including understanding the diet and mobility of Bronze Age populations. Although it is important to acknowledge the fundamentally continuous nature of societal development, the Bronze Age in Czechia is conventionally divided into three main chronological phases, each defined by markedly different archaeological patterns—particularly in funerary practices. The Early Bronze Age (EBA) is represented by the Únˇ etice culture, which is characterized by extensive inhumation cemeteries. The Middle Bronze Age (MBA), associated with the Tumulus culture, is defined by the emergence of burial mounds. The Late and Final Bronze Age (LBA/FBA), in contrast, is dominated by large urnfields containing hundreds to thousands of cremation burials, although occasional inhumation remains also appear in settlement pits or at hilltop sites (Harding 2000; Jir´ aˇ n 2013). These pronounced shifts in mortuary practices may reflect broader changes in the behaviour, social structure, and worldview of settled populations—changes that can also be explored through isotopic analyses of human skeletal remains, offering insights into aspects such as mobility, diet, and regional interaction. 1.1. Stable isotope analysis in the dietary reconstruction of past populations Stable isotope analysis refines the analysis of past diets by deriving the individual dietary signal from the mineralized human tissues (LeeThorp 2008). Following the rule that “you are what you eat”, the isotopic signature of each organism’s tissues reflects their diet. The combination of carbon (ẟ 13 C) and nitrogen (ẟ 15 N) isotopic values makes it possible to identify the dietary proportion of basic isotopically distinct food groups. Carbon isotopic values help to distinguish between marine and terrestrial foods and/or diets based on plants with different photosynthetic pathways for CO 2 fixation. In general, marine organisms and plants of the so-called C4 photosynthetic pathway (such as maize or millet) have significantly higher ẟ 13 C values than terrestrial C3 plants (Ambrose and Norr 1993; DeNiro and Epstein 1978). Both ẟ 13 C and especially ẟ 15 N values also reflect the trophic position of the organism in the local food web. It has been shown that the bone collagen of herbivores has δ 13 C values about 5 ‰ higher than those of their feed. (Ambrose et al. 1997). Higher up the food chain, the difference in collagen δ 13 C values between prey and consumer is about 1 ‰ (Bocherens and Drucker 2003; Lee-Thorp 2008). For δ 15 N values a much higher trophic level step has been observed, ranging from 3-6 ‰, based on controlled feeding experiments with different animal species and direct observation for humans (Fernandes et al. 2012; Hedges and Reynard 2007; O’Connell et al. 2012). Due to the generally longer food chains (and higher trophic level effects), foods of aquatic origin (both marine and freshwater) tend to have significantly higher ẟ 15 N values than terrestrial foods. In the freshwater environment, however, both carbon and nitrogen isotope values are highly variable. (Dufour et al. 1999; Katzenberg and Weber 1999). For the purposes of this study, ẟ 13 C and ẟ 15 N values were measured in bone collagen, which primarily reflects the isotopic signature of the protein consumed, rather than that of the whole diet. However, the use of a Bayesian model of food reconstruction using isotopically transferred signals (FRUITS; Fernandes et al. 2014) partially filters out this issue, and allows different sources of uncertainty and offsets between diet and tissue to be considered when modelling the relative contribution of isotopically distinct food groups. 1.2. Current knowledge on Bronze Age agriculture In the context of Central Europe, the Bronze Age is characterised by a predominantly warm climate with less precipitation than the preceding Neolithic period (Poschlod 2015). The systematic analysis of archaeobotanical remains on Czech territory has revealed a significant change in the composition of cultivated plants between the Early Bronze Age (EBA; 2200–1500 BCE) and the Middle Bronze Age (MBA; 1500–1300 BCE; Hajnalov´ a 2012). During the Early Bronze Age, a poor range of field crops survived from the Eneolithic, with a clear dominance of emmer, accompanied mainly by barley and einkorn (Opravil 1996). Legumes are scarce, being represented mainly by garden pea and lentil (Koˇ c´ ar and Dreslerov´ a 2010). During the Middle Bronze Age, in addition to the significant introduction of spelt, the main novelty was the introduction of broomcorn millet (Panicum miliaceum; Dreslerov´ a and Koˇ c´ ar 2013), a plant of the C4 photosynthetic pathway. It was not before the Middle Bronze Age that millet grains appeared in significant quantities in the Czech territory, including grain accumulations (depots; Bernardov´ a 2009; Dreslerov´ a and Koˇ c´ ar 2013). During the MBA, millet became dominant, or at least subdominant on many sites (Pokorn´ a et al. 2024). The sudden emergence and unprecedented popularity of millet is consistent with the broader European pattern, as is currently being gradually reconstructed through the combined use of archaeobotany, isotopic analysis and radiocarbon dating. Millet was domesticated in China between 8000 and 6000 BCE (Stevens et al. 2021; Yang et al. 2012) and spread westwards via Central Asia, where it arrived between 2300 and 2000 BCE (Frachetti et al. 2010). Its further progress towards Europe is not fully documented, but it was probably via the Caucasus or Turkey (Martin et al. 2021; Motuzait´ e-Matuzevici¯ ut˙ e et al. 2013), with the oldest European finds of millet coming from the area of Ukraine in the 17th-16th centuries BCE (Dal Corso et al. 2022; Filipovi´ c et al. 2020). The middle of the second millennium (1600–1400 BCE) is a period of the sudden, one might even say explosive, expansion of millet in most of western Eurasia (e.g. Bouby et al. 1999; Jantzen et al. 2011; Pospiszny et al. 2021; Tafuri et al. 2009). Finally, around 1000 BCE, millet reached the edge of its geographical limits in the area of northeastern Europe (Motuzait´ e-Matuzevici¯ ut˙ e and Lauˇ zikas 2023).Back in Czech territory, while barley and emmer wheat remained staple crops during the Late Bronze Age (LBA), this period—dated to approximately 1300/1250–1050/950 BCE, depending on regional chronological schemes in Moravia and Bohemia—also marks the peak in the occurrence of millet. At the same time, lentils became the dominant legume, accompanied by smaller quantities of garden pea, faba bean, and common vetch. In the Final Bronze Age (1050/950–800 BCE), the importance of millet declined.. The MBA and LBA sites were highly diversified, suggesting that the crops were selected to fit local conditions (Pokorn´ a et al. 2024; ˇ S´ alkov´ a et al. 2019). The results of archaeozoological analyses suggest that people consumed meat and other products mainly from domestic animals, while hunting clearly played only a supplementary role (Roblíˇ ckov´ a 2003). The distribution of slaughter ages suggests that domestic animals were bred not only for meat but possibly also for milk, labour, hair, or wool (Kovaˇ cikov´ a and Troj´ ankov´ a 2018). Osteological evidence from the Early Bronze Age indicates that cattle were predominant, followed by caprines, and domestic pigs, while the prevalence of small ungulates varied with local environmental conditions (Hl´ asek et al. 2023). The same tendency is evident in MBA and LBA settlements, although in some cases pig bones increased at the expense of caprines (Roblíˇ ckov´ a 2003). Occasionally, other animals were hunted or collected as food supplements, including wild mammals, turtles, fish (mainly cyprinids), and freshwater bivalves, such as the thick-shelled river mussel or the painter’s mussel (Kovaˇ cikov´ a and Troj´ ankov´ a 2018; ˇ Siroký et al. 2004). 1.3. Aims of the study in the context of the Central European Bronze Age Compared to archaeobotanical and archaeozoological evidence, the isotopic approach to reconstructing Bronze Age diets has been rather underrepresented, focusing solely on case studies reconstructing individual dietary histories (Kaupov´ a et al. 2019b; Le Huray and Schutkowski 2005; Parma 2017; Parma et al. 2018; Salaˇ s et al. 2012a; 2012b). S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 2 This has mainly been due to the difficulties in assembling representative datasets because of the preferential burial practices in the later phases of the Bronze Age. In the EBA inhumation was almost exclusively favoured, with a predominance of small group burials, but with some depositions of human remains also found in settlement contexts. In the MBA cremation began to appear on a large scale, and became exclusive in the LBA. During the MBA and LBA inhumations are found in settlement contexts exclusively in two possible scenarios: the first, less numerous, where they are found in the context of fortifications in a few central locations (Harding and Palmer 2007), and the second, more common, deposition of human remains in settlement pits. This phenomenon clearly limits the application of carbon and nitrogen stable isotope analysis to only a small part of the MBA and LBA population. However, with this limitation in mind, isotopic analysis might still be able to identify the main dietary shifts during the Bronze Age. Within this framework, the aim of this study was to obtain the first direct and quantitative information on millet consumption in LBA populations. Given that crop and domestic animal diversification likely reflects adaptive responses by MBA and LBA communities to local environmental conditions (Harding 2000; Pokorn´ a et al. 2024; ˇ S´ alkov´ a et al. 2019), the diversity of subsistence patterns will be examined not only between but also within sites. Furthermore, as the proportion of animal products in the diet is an important indicator of social status across time and space (Kinpper et al. 2015), variation in nitrogen isotope values can be considered an indirect indicator of the degree of social stratification in EBA populations. In this context, it should be noted that a comparison of the anthropological parameters (such as stature) of noncremated LBA individuals with those from other periods suggests a potential disadvantage for this population group—excluded from the regular burial rite of cremation—in terms of access to high-quality food, such as nutritionally rich animal products (Dobisíkov´ a et al. 2007). Therefore, this study will focus on the diet as a reflection of the living conditions of the LBA minority, which was exempted from the prevailing burial rite. Finally, we will investigate potential dietary variation at the studied sites with respect to burial context (ritual vs. non-ritual, cemetery vs. settlement burials). 2. Material The total of 103 human samples, 30 animal samples (cattle, sheep/ goat, domestic pig, dog, red deer, and aurochs) and 8 archaeobotanical samples (einkorn, emmer, spelt, barley, and millet) analysed for this study (Table 1) were divided into 4 subgroups according to their geographical location (Bohemia vs Moravia) and chronology (EBA vs LBA; Fig. 1). Regrettably, non-cremated skeletal material from the MBA is very scarce, mostly badly preserved, and was not accessible at the time of the study. Bohemia and Moravia are two historical regions of what is now the Czech Republic. Although this division originates in the Middle Ages, the populations of these regions often exhibited distinct patterns of development throughout both the prehistoric and historical periods. During the Neolithic and the Bronze Age, human settlements were concentrated in the most fertile areas along rivers. The regions are separated by the Bohemian-Moravian Highlands; although the highlands did not constitute a true geographical barrier, settlement in this area was significantly sparser. As a result, the highlands may have functioned as a filter or delaying factor for migration and cultural exchange between the western and eastern regions throughout much of prehistory (Hejhal 2009; Jir´ aˇ n 2013). Distinct patterns of development were already apparent in the Neolithic period (e.g. the Stroked Pottery Culture in Bohemia versus the Moravian Painted Ware culture in Moravia; Pavlů 2004). During the Bronze Age, these differences became more pronounced, with Moravia maintaining connections to the Danube region (Slovakia, Hungary), while Bohemia was more closely linked to Central Germany. In the EBA, both regions are represented by the Únˇ etice culture, but in slightly different variants (e.g. in ceramic styles). During the Middle Bronze Age, the Tumulus Culture was typical of Bohemia and South Moravia, while in Central Moravia the Lusatian Urnfield culture began to crystalise. In the Late Bronze Age, the Knovíz Culture appeared in most of Bohemia, accompagnied by the Lusatian Urnfield Culture in its eastern part. In South Moravia, the Middle Danube Urnfield culture is represented by the Velatice phase, while North Moravia was inhabited by people of the Lusatian culture. (Jir´ aˇ n 2013; Podborský 1993). It should be noted that no significant differences have been observed between the two regions in the composition of archaeozoological assemblages. The distribution of plant macroremain assemblages reveals several spatial trends that appear to be more closely linked to local environmental factors—such as altitude, soil quality, and precipitation levels—than to broader geographical divisions like Bohemia and Moravia (Pokorn´ a et al. 2024). However, the fact that Moravia has, on average, a lower altitude than Bohemia results in some differences in the relative abundance of wheat and barley. Most importantly, in this context, millet was dominant in lowland areas during the Late Bronze Age, regardless of soil quality (Dreslerov´ a et al. 2017). EBA Bohemia is represented by a sample from the East Bohemian cemetery at Mikulovice, an important junction on a long-distance trade route, the so-called Amber Road (Ern´ ee and Langov´ a 2020). Only adult individuals with estimated sex were selected for stable isotope analysis, resulting in a sample of 44 individuals. The isotopic values of these individuals have already been published in Ern´ ee and Langov´ a (2020), but this is the first time they have been presented in a broader chronogeographic conetext. The region of Moravia is represented by the Early Bronze Age Únˇ etice culture site at Brno-Tuˇ rany, where 13 burials in settlement pits were uncovered in an extensive settlement area. The adjacent cemetery contained 28 individuals (Moravcov´ a and Kala 2019). A representative subsample of individuals from the settlement pits (n =6) and cemetery (n =15) was selected for analysis, including adults and subadults over 2 years of age. LBA Bohemia is represented by the Central Bohemian site of Z´ alezlice. All the skeletal material was excavated from the settlement Table 1 Description of studied samples. Site Location Period Dating (BCE) Humans (n) Animals (n) Plants (n) Mikulovice Bohemia EBA 2200/2100–1800/1700 44 7: C(2), S/G(2), P(2), RD (1) 0 Z´ alezlice Bohemia LBA 1100–950 8 8: C(1), S/G (2), P(2), D (2), A(1) 0 Plaˇ cice Bohemia EBA 0 0 2: EM(2) Bavorynˇ e Bohemia LBA 0 0 3: EM(1), B (1), M(1) Brno-Tuˇ rany Moravia EBA 2000–1700 21 11: C(3), S/G (6), P (2) 0 Bluˇ cina Cezavy Moravia LBA 1300–1050 25 4: C(1), S/G (1), P (2) 3: EI (1); S (1), M(1) Hoˇ stice Moravia LBA 1300–1050 4 0 0 Ivanovice na Han´ e Moravia LBA 1050–900 1 0 0 Vyˇ skov Moravia LBA 1300–1050 1 0 0 C =cattle, S/G =sheep/goat, P =domestic pig, D =dog, RD =red deer, A =aurochs, EI =einkorn, EM =emmer, B =barley, M =millet, S =spelt; No of samples in parentheses. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 3 area, giving a total of 8 adults and juveniles over 11 years of age (Unger and Pecinovsk´ a 2015). The Late Bronze Age population of Moravia is represented mostly by the Bluˇ cina Cezavy hillfort site, providing the skeletal remains of 25 adults and subadults over 6 years of age that could reliably be dated to the LBA period (Salaˇ s 2023). Four of these individuals were previously analysed and published by Drtikolov´ a Kaupov´ a et al. (2019b). The LBA dataset from Moravia was supplemented by several samples from other contexts. Three of these come from the Hoˇ stice cemetery; of the 21 burials recovered, only 3 were not cremated and could therefore be included in the study (Parma and Stuchlík 2017). Next, a female skeleton from Ivanovice na Han´ e (Parma et al. 2018) was deposited in a secluded storage pit not associated with any other signs of settlement activity, the isotopic values have already been published (Parma et al. 2018). Radiocarbon dating of this skeleton (Table 1) places the individual at the transition between the Late and Final Bronze Age (LBA/ FBA). However, for the sake of simplicity, we have included it in the LBA category. Finally, a subadult individual from Vyˇ skov-Nouzka was deposited non-ritually in the settlement area in a silo pit with a collection of artefacts that allowed dating (Parma et al. 2014). The animal and human bone material was taken from the respective archaeological excavations. In the case of LBA Moravia, only animals from Bluˇ cina Cezavy were analysed. Fish bones from selected sites were not available for analysis. However, as the anthropogenic influence on fish isotope values was probably minimal before the development of pond management in the High Middle Ages (ˇ Sarapatka et al. 2014), previously published data for Early Medieval Central Europe (Kaupov´ a et al. 2018; 2019a; Reitsema et al. 2013) were used. Apart from Bluˇ cina Cezavy, no plant macroremains were collected at the investigated sites. The archaeobotanical dataset was completed with samples from Plaˇ cice (EBA Bohemia) and Bavorynˇ e (LBA Bohemia). From the description of the studied sites, it is clear that our dataset is not representative of each chrono-geographical context. Rather, it reflects the preservation and accessibility of skeletal material, and should therefore be considered a pilot study. More details of the studied sites are given in Table 1 and Online supplementary material 1. 3. Methods The extraction of collagen from bone and dentine samples and the pre-treatment of plant samples were carried out in the laboratory of the Department of Anthropology of the National Museum (Prague, Czech Republic). Samples were taken preferentially from the cortical bone of ribs. In cases of poor preservation, long bones of the upper and lower limbs were sampled. Dentine was analysed only in cases of extremely poor preservation of cortical bone (4 individuals from Bluˇ cina Cezavy, 1 from Hoˇ stice and 1 from Vyˇ skov); in these cases, differences in the life span reflected in the isotopic signal of bone and dentine were taken into account in the data interpretation (Balasse et al. 2001). In order to avoid the breastfeeding signal (Fuller et al. 2006), the roots of second molars were preferentially analysed, reflecting diet at the age of approximately 8.5–13.5 years (AlQahtani et al. 2010). Due to the young age of the Vyˇ skov individual, the root of a first primary molar (74) was sampled, retaining an isotopic signal from approximately 7.5 months to 2.5 years Fig. 1. Geographical location of the sites included in this study: A– EBA sites, B – LBA sites, C – Bohemia, D – Moravia. Numbering of the sites: 1 – Mikulovice; 2 – Brno-Tuˇ rany; 3 – Z´ alezlice; 4 – Bluˇ cina Cezavy; 5 – Hoˇ stice; 6 – Ivanovice na Han´ e; 7 −Vyˇ skov. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 4 of age (AlQahtani et al. 2010). This individual was included to observe their carbon isotopic values, with the effect of significant millet consumption expected to clearly overwhelm the slight trophic level effect of breastfeeding. A minimum of 200 mg of cortical bone or 50 mg of dentine was collected from each individual. Collagen was extracted according to the Longin (1971) method as modified by Bocherens (1992). Charred grains (10 grains per sample) were rinsed with deionised water and homogenised using an agate mortar and pestle. No further pretreatment was performed, as previous studies (Brinkkemper et al. 2018; Lightfoot and Stevens 2012) suggest that there is no difference between pretreated and untreated samples, even in cases where sample contamination was suggested by FTIR analysis. Elemental Analysis – Isotope Ratio Mass Spectrometry (EA-IMRS) was performed at Iso-Analytical, Crewe, United Kingdom. Details on analytical precision are given in Online supplementary material 2. Quantitative diet reconstruction was performed using Bayesian mixture modelling in FRUITS 3.0.0. software (Fernandes et al. 2014). The model was weighted using relative nutrient intake estimates constrained by physiological limits of protein consumption. As we are aware of the weaknesses of the model (Cheung and Szpak 2022) arising from 1) the small sample sizes for some food groups (namely plants) as well as for some population groups (namely LBA Bohemia) and 2) the possible overlap in isotopic values between C3 plants and terrestrial animals (Fig. 2), we present here only the estimated dietary intake of millet as a first quantitative estimate, with a view to improving the quality of the model in the future. Thus, for the time being, we deliberately do not present and discuss other −more subtle −variations observed for other food groups. The full specification of the model parameters, as well as the complete FRUITS output, can be found in Online supplementary material 3. Isotopic data were compared using an independent samples t-test or ANOVA (with post hoc Tukey HSD), or a non-parametric Mann–Whitney test (the exact variant for small samples) with respect to biological characteristics (sex and age) and archaeological characteristics (type of burial), as well as chrono-geographical categorisation. When Levene’s test indicated significant differences in variances, Welch’s ANOVA was used instead of the standard ANOVA, followed by the Games-Howell post hoc test.Statistical analysis was performed in RStudio 1.2.5033. 4. Results Full isotopic data are available in Online supplementary file 4 and on the Isoarch platform (https://doi.org/10.48530/isoarch.2024.007; Plomp et al. 2022; Salesse et al. 2018). The C:N ratios of the grain samples ranged from 19.6 to 25.2, with a median of 21.9. The percentage of carbon ranged from 31.9 % to 62.5 % with a median of 46.7 %, while the percentage of nitrogen ranged from 1.9 % to 3.3 % with a median of 2.4 %. The ẟ 13 C values of the C3 plant grain samples (n =6) ranged from −24.4 ‰ to –22.6 ‰, with a median of –22.9 ‰. After adjustment for the charring effect (by subtracting 1 ‰; Fraser et al. 2013), their ẟ 15 N values ranged from 4.2 ‰ to 7.1 ‰ with a median of 4.7 ‰. The ẟ 13 C values of the two millet samples were almost identical, being −10.2 ‰ for the Plaˇ cice and −10.5 ‰ for the Bluˇ cina Cezavy sample, while their ẟ 15 N values were 6.2 ‰ and 5.4 ‰ respectively (Figs. 2 and 3; Online supplementary material 4). Due to the small sample size, we did not attempt to divide the plant dataset according to the chronogeographic context Drtikolova_2024_Fig_2.jpg Fig. 2. Complete isotopic data from Bronze Age sites in Czechia; for animal samples; isotopic values of wild herbivores and dogs are not shown. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 5 for any statistical analysis. All the collagen samples met the preservation criteria (Ambrose 1990; DeNiro 1985; Van Klinken 1999; Online supplementary material 4). From the animal dataset, two samples of wild fauna showed outlying isotopic values – red deer (ẟ 13 C =–23.1 ‰; ẟ 15 N =5. 7 ‰) and aurochs (ẟ 13 C =–22.0 ‰; ẟ 15 N =5.0 ‰), – as did two dog samples (ẟ 13 C = −18.7 ‰, −18.4 ‰; ẟ 15 N =9.7 ‰, 9.1 ‰; Fig. 3). Excluding these samples, ẟ 13 C values of domesticated herbivores and pigs ranged from –23.2 ‰ to −19.3 ‰ with a median of −20.8 ‰, while ẟ 15 N values ranged from 6.8 ‰ to 10.7 ‰ with a median of 8.1 ‰. Domesticated herbivores (N =18) showed ẟ 13 C values from –23.2 ‰ to −19.4 ‰ with a median of −20.7 ‰ and ẟ 15 N values from 6.8 ‰ to 10.7 ‰ with a median of 7.8 ‰, while pigs (N =8) showed ẟ 13 C values from –22.2 ‰ to −19.3 ‰ with a median of −21.1 ‰ and ẟ 15 N values from 7.2 ‰ to 9.7 ‰ with a median of 8.2 ‰. In the reduced data set of domesticated herbivores and pigs, there were no statistically significant differences between chronogeographic contexts in either ẟ 13 C (Kruskal-Wallis test; p =0.056) or ẟ 15 N (KruskalWallis test; p =0.692) values. However, in the case of carbon isotopic values, the result was close to the 0.05 significance level, with the EBA Bohemia sample showing lower ẟ 13 C values than the other groups (Fig. 3, Table 2). For the following analysis of human diet, including FRUITS modelling, a combined sample of domesticated herbivores and pigs was used, regardless of the geographical or chronological context. However, the suspected slight variation in ẟ 13 C values is taken into account when discussing human data. Human ẟ 13 C values in EBA Bohemia ranged from −21.0 ‰ to −19.8 ‰ with a median of −20.4 ‰ while ẟ 15 N values ranged from 10.6 ‰ to 13.4 ‰ with a median of 11.2 ‰ (Table 3). In this subsample, females (median =-20.3 ‰) had higher carbon values than males (median = -20.5 ‰; Table 4). The average calorie contribution of C4 plants as estimated by the FRUITS model was 5.0 ±4.5 %. In EBA Moravia, ẟ 13 C values ranged from −20.2 ‰ to −18.9 ‰ with a median of −19.6 ‰ and ẟ 15 N isotopic values ranged from 9.7 ‰ to 13.3 ‰ with a median of 11.4 ‰. There were no statistically significant differences based on sex, age or type of burial (burial site vs. settlement pit; Table 4). The average calorie contribution of C4 plants was 6.8 ± 5.2 %. In LBA Bohemia ẟ 13 C values ranged from −18.4 ‰ to −16.1 ‰ with a median of −16.9 ‰ and ẟ 15 N values ranged from 9.6 ‰ to 10.7 ‰ with a median of 10.2 ‰. Although none of the statistical operations performed on such a small dataset have high power, for the sake of completeness it must be said that there were no statistically significant differences based on age or type of burial (settlement ritual vs. settlement non-ritual). As there was only one male in the LBA Bohemia sample, the effect of sex could not be evaluated. The average calorie contribution of C4 plants Drtikolova_2024_Fig_3.jpg Fig. 3. Detailed view of the plant and animal isotopic data from Bronze Age Czechia; isotopic values of millet samples are not shown. Table 2 Mean carbon and nitrogen isotopic values in animal samples (dogs, red deer and auroch excluded) with respect to chronogeographic context. δ 13 Cδ 15 N n Median Mean SD Median Mean SD EBA Bohemia 6 −21.5 −21.4 0.4 7.6 7.9 1.1 EBA Moravia 11 −20.7 −20.8 1.1 8.6 8.6 1.5 LBA Bohemia 5 −20.3 −20.3 0.5 7.7 7.8 0.7 LBA Moravia 4 −20.9 −20.9 0.3 8.3 8.5 0.9 EBA =Early Bronze Age, LBA =Late Bronze Age, n =number of individuals, SD =standard deviation S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 6 was 24.0 ±8.8 %. In LBA Moravia ẟ 13 C values ranged from −21.1 ‰ to −13.1 ‰ with no statistically significant differences by sex or age. When estimating the effect of age, adult individuals from whom dentine samples were taken were treated as subadults. Since in the case of the Bluˇ cina Cezavy site it is not possible to speak of burials in the strict sense, the effect of burial type cannot be evaluated. The average calorie contribution of C4 plants was 34.4 ±9.2 %. When testing for dietary differences between the four chronogeographical contexts, ANOVA revealed statistically significant differences in both ẟ 13 C and ẟ 15 N values (Table 5). Carbon isotope values were statistically different between all the investigated subsamples, with LBA samples showing far more positive ẟ 13 C values than EBA in both geographical contexts (Table 5, Figs. 2 and 4). At the same time, Moravian samples from both periods showed slightly more positive δ 13 C values than those from Bohemia. Due to the aforementioned variation in δ 13 C values observed in the animal dataset, statistical operations were repeated comparing human-faunal isotopic offsets counted separately for each time period and geographical context. The results of the statistical operations mostly did not differ from those observed on the raw Table 3 Basic statistics of human isotopic data. Context ẟ 13 C (‰)ẟ 15 N (‰) n Median Mean ± SD Δ 13 C h-fa Median Mean ± SD Δ 15 N h-fa EBA Bohemia All (adults) 44 −20.4 −20.4 ±0.3 0.5 11.2 11.3 ±0.6 3.0 EBA Moravia All 21 −19.6 −19.6 ±0.3 1.3 11.4 11.4 ±0.9 3.1 adults 15 −19.6 −19.6 ±0.3 1.2 11.3 11.2 ±0.8 2.9 non-adults 6−19.5 −19.4 ±0.3 1.5 12.0 11.9 ±1.1 3.7 LBA Bohemia All 8−16.9 −17.0 ±0.7 3.9 10.2 10.2 ±0.4 1.9 adults 5−17.0 −16.9 ±0.5 4.0 10.4 10.3 ±0.4 2.0 non-adults 3−16.8 −17.2 ±1.1 3.7 10.0 10.0 ±0.5 1.8 LBA Moravia All 31 −15.6 −15.9 ±1.5 5.0 10.5 11.6 ±1.2 2.4 adults 18 −15.6 −15.7 ±0.8 5.2 10.8 10.8 ±0.5 2.5 non-adults 13 −15.6 −16.1 ±2.2 4.8 10.3 10.4 ±1.7 2.1 a human-faunal isotopic offset; as faunal data, the mean values of the main consumed domesticated animals (cattle, sheep/goat, domestic pig) were used. Table 4 The impact of sex, age and type of burial (BS =burial sites; SR =ritual burial in settlement area; SNR =non-ritual burial in the settlement area) on carbon and nitrogen isotopic values. n p ẟ 13 Cẟ 15 N EBA Bohemia Sex (Male/Female) 20/24 0.006 t 0.503 t   Age (Adult/Non-adult) 44/0 x x   Burial type (BS/SR/SNR) 44/0/0 x x EBA Moravia Sex (Male/Female) 6/9 0.607 M 0.224 M   Age (Adult/Non-adult) 15/6 0.132 M 0.235 M   Burial type (BS/SR/SNR) 15/6/0 0.677 M 1.000 M LBA Bohemia Sex (Male/Female) 3/1 x x   Age (Adult/Non-adult) 5/3 1.000 M 0.571 M   Burial type (BS/SR/SNR) 0/5/3 0.786 M 0.571 M LBA Moravia Sex (Male/Female) 5/9 1.000 M 0.147 M   Age (Adult/Non-adult) 8/13 1.000 M 0.087 M   Burial type (BS/SR/SNR) 3/2/0 x x M =Mann-Whithey test (exact variant for the small sample sizes); t =independent sample t-test; results significant at 0.05 level are in bold. Table 5 Comparison of stable isotope values of Early and Late Bronze Age Bohemia and Moravia subsamples. All groups compared with ANOVA/Welch’s ANOVA, intergroup comparison with post-hocTukey HSD/Games-Howell test, significant results in bold. ẟ 13 C a (Δ 13 C h-f ) a ẟ 15 N (adults only) All <0.001 (<0.001) <0.001 (0.006) EBA Bohemia vs EBA Moravia 0.002 (0.745) 0.965 (0.963) EBA Bohemia vs LBA Bohemia <0.001 (<0.001) 0.008 (0.045) EBA Bohemia vs LBA Moravia <0.001 (<0.001) 0.009 (0.026) EBA Moravia vs LBA Moravia <0.001 (<0.001) 0.012 (0.261) EBA Moravia vs LBA Bohemia <0.001 (<0.001) 0.007 (0.141) LBA Moravia vs LBA Bohemia 0.011 (<0.001) 0.595 (0.716) a as the data did not meet the condition of homogeneity of variances, Welch’s ANOVA with a post-hoc Games-Howell test was used. Fig. 4. The violin plots illustrate the full distribution and density of the δ 13 C (A) and δ1 5 N isotope values within each chronogeographical context. Embedded within each violin plot, the boxplots display the median, interquartile range, and range excluding outliers, providing a summary of the central tendency and variability. Individual data points represent single samples and are color-coded according to the defined age groups, allowing visualization of within-group variability. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 7 isotopic data (Table 5), except for the comparison of EBA samples in Moravia vs. Bohemia, where the result lost statistical significance. Nitrogen isotopic values were significantly lower in the LBA than in the EBA in both geographical contexts. There was no statistically significant difference between Moravian and Bohemian samples in any time period. As the effect of breastfeeding on ẟ 15 N values (Fuller et al. 2006) cannot be excluded for some of the subadults, statistical operations for ẟ 15 N values were repeated in the reduced sample of adults. This did not change most of the results, but δ 15 N values from the EBA Moravia context were no higher compared to the LBA. (Table 5). 5. Discussion 5.1. Plant and faunal isotopic data To adjust the δ 13 C values of grain samples for the temporal variation in the isotopic composition of atmospheric CO 2 (Ferrio et al. 2003), a correction factor of 6.45 ‰ was applied as an average for the period 2000 to 800 BCE. The difference in δ 13 C values of atmospheric CO 2 between the Early and Late Bronze Age appeared to be less than 0.1 ‰ and was therefore omitted. For wheat grain (different species combined), Δ 13 C values ranged between 16.2 ‰ and 16.8 ‰. A sample of barley has a Δ 13 C value of 18.1 ‰. This in both cases corresponds to moderately watered plants (Wallace et al. 2013). This is similar to other prehistoric sites from Central Europe, ranging from Neolithic Germany (Bogaard et al. 2013) via Iron Age Switzerland (Knipper et al. 2015) to Early Medieval Czechia (L´ atkov´ a et al. In Press). Carbon isotopic values of two millet samples are also comparable with prehistoric Polish samples (Mueller-Bieniek et al. 2019). Nitrogen isotopic values of grains mainly correspond to medium manuring rates, but in two samples from EBA Bohemia, a high rate of manuring may be suggested (Bogaard et al. 2013). Although it would be premature to generalise from a data set of this size, the current plant data suggest the following: firstly, there is no trace of poor water availability, even though in the Central European context, the Bronze Age climate is characterised as relatively hot and dry (Poschold 2015). Secondly, there may have been a change in agricultural practices, such as manuring, affecting cereal δ 15 N values between the Early and Late Bronze Age, which has previously been suggested by archaeobotanists (Hajnalov´ a 2012). However, the latter finding in particular should be seen as a starting point for further research and as a potentially valid factor in the discussion of human isotopic data, rather than a conclusion of any kind. With the exception of two LBA dogs, terrestrial faunal δ 13 C values were characteristic of the C3 plant environment. As there is an enrichment in 13 C of about 5 ‰ between the collagen of herbivores and their food (Ambrose et al. 1997), the δ 13 C values of domesticated animal forage varied around −25.8 ‰. Returning to plants, δ 13 C values of wheat and barley chaff should average between 23.9 and 24.9 ‰ after adjustment for an isotopic offset between grain and chaff/leaves (Merah et al. 2002). Herbivore forage was thus somewhat depleted in 13 C compared to crop chaff, suggesting grazing in wetter and/or more shaded habitats (Arens et al. 2000; Flohr et al. 2019). The lower δ 13 C values of the EBA Bohemia faunal sample in comparison to Moravian contexts is consistent with a trend observed for Czech territory in other time periods (Drtikolov´ a Kaupov´ a et al. 2023b; Kaupov´ a et al. 2019a). However, the LBA Bohemia sample diverges from this picture. The nitrogen isotopic values of domesticated herbivores and pigs are comparable between the EBA and LBA contexts and are, on average, somewhat higher than those from most other chronological contexts on Czech territory (Drtikolov´ a Kaupov´ a et al. 2023b; Vytlaˇ cil et al. 2024). At the same time, the δ 15 N values of C3 cereals are only 0.9 ‰ higher than those estimated for animal fodder (mean δ 15 N =4.1 ‰), based on an estimated trophic level shift of 3–5 ‰ (Hedges and Reynard 2007). The relatively high δ 15 N values in forages may be caused by intensive grazing, seasonal grazing on manured land and/or the seasonal spreading of agricultural products from manured fields (Reitsema et al. 2013). It should be noted that the δ 15 N values of sheep/goats in particular are highly variable with some extreme values of almost 11 ‰ (Fig. 3). The two dog samples from the Z´ alezlice LBAB site have distinct isotopic values (Fig. 3), reflecting both their carnivorous nutritional strategy and their close coexistence with humans (Guiry 2012). The two samples of wild fauna (EBAB sample of red deer MKVF04 and LBAB sample of aurochs ZALF10) both show δ 13 C values lower than –22 ‰, consistent with a forested habitat (Hofman-Kami´ nska et al. 2018). The δ 15 N values of both individuals are lower than those observed in any of the domesticated herbivores, suggesting that the markedly elevated δ 15 N values of livestock are due to human husbandry practices rather than the general environmental context of the period studied. 5.2. Human dietary habits Statistically significant differences for both δ 13 C and δ 15 N values observed between the defined chrono-cultural contexts attest to notable changes in diet and/or agricultural practices (Figs. 2 and 4). In both EBA contexts, the low average Δ13C human-fauna (0.5 ‰ in Bohemia; 1.3 ‰ in Moravia) and the high homogeneity of the δ 13 C values (1SD =0.3 ‰ for both Moravia and Bohemia) reflect a terrestrial diet without a significant input of C4 plants. This is further confirmed by the FRUITS modelling, which shows that the average dietary contribution of C4 plants varies between 5 % and 7 % in EBA contexts with minimum estimates approaching 0 %. At this point, it is important to remember that FRUITS cannot calculate the mean dietary contribution of the respective food group to be zero, and that an estimated C4 contribution of around 10 % is common even in contexts with no documented presence of C4 plants (Zavodny et al. 2017). The fact that the small but notable difference in δ 13 C values between EBA Moravia and Bohemia lost its statistical significance when comparing human-faunal isotopic offsets suggests that this most likely reflects isotopic differences at lower levels of the food chain, rather than dietary differences. The remarkable input of millet is clearly documented in both LBA contexts, with Δ 13 C human-fauna at least double the threshold value generally considered indicative of notable millet consumption (2 ‰; Lightfoot et al. 2012; Table 3). Such a high proportion of millet in the diet has not previously been documented in Czech prehistoric and historic contexts (Drtikolov´ a Kaupov´ a et al. 2023a; Kaupov´ a et al. 2018; 2019a; Le Huray and Schutkowski 2005; Plecerov´ a et al. 2020; Salesse et al. 2013; Vytlaˇ cil et al. 2024). Comparing the two geographical contexts, the Moravian sample showed a higher average millet intake (34 %) than the Bohemian sample (24 %). As the statistically significant difference between the two LBA geographical units persists when the variation in faunal values described above is taken into account, this clearly demonstrates that there was considerable variation in millet consumption across Czech territory. However, since the Bohemian context is represented by only one site, and the Moravian sample is clearly dominated by the Bluˇ cina Cezavy site, it may be misleading to directly relate the variation described above to the main geographical unit. Local climatic conditions or the hillfort character of the Bluˇ cina Cezavy site may well have contributed to the observed pattern. Nitrogen isotopic values indicate that food of terrestrial origin dominated the diet of Bronze Age individuals. The proportion of animal products probably differed significantly between the chronogeographical contexts. In both EBA contexts the Δ 15 N human-fauna (3.0 ‰ in Bohemia; 3.1 ‰ in Moravia) was significantly higher than in the LBA (1.9 ‰ in Bohemia; 2.4 ‰ in Moravia). However, with the current sample size, alternative explanations for the observed temporal pattern cannot be excluded. The present animal data do not indicate a change in the composition or properties of the S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 8 feed ingested over time. Although no statistical comparison was made due to the small sample size, the isotopic values of pigs appear to be similar to those of domesticated herbivores (Fig. 3). It is therefore unlikely that a higher consumption of pork in LBA populations could have caused the observed isotopic shift. A more plausible explanation may lie in a change in the composition of the plant component of the diet, as legumes could cause a decrease in the δ 15 N values of the consumer (DeNiro and Epstein 1981). Although the importance of legumes is difficult to estimate due to their different taphonomy, current archaeobotanical data suggest that the proportion of legumes increased in the LBA (Koˇ c´ ar and Dreslerov´ a 2010; Pokorn´ a et al. 2024; ˇ S´ alkov´ a et al. 2019). Finally, a possible change in agricultural practices affecting cereal δ 15 N values between the EBA and LBA, such as different manuring intensity, has been suggested by previous archaeobotanical findings (Dreslerov´ a et al. 2017; Hajnalov´ a 2012; Pokorn´ a et al. 2024). The increasing diversity of crops and their widespread cultivation may reflect population growth, rising food demand, or growing pressure on arable land, supposedly leading to the intensification of agricultural practices such as manuring during the Bronze Age. An increased level of manuring during this period has been documented in other European regions (e.g. Varalli et al. 2021); however, our small plant dataset—although it must be interpreted with caution—does not support this scenario, as both human and plant isotopic values within it decrease between the Early and Late Bronze Age.. Despite the above-outlined alternatives, the relatively low δ 15 N values observed in LBA Bohemia may indicate a certain restriction of the high-quality diet in individuals exempted from the regular burial treatment of cremation. This is in line with their anthropological characteristics, namely their small average stature (Dobisíkov´ a et al. 2007). Alternatively, the strategic location of the Mikulovice EBA site at a junction on the Amber Road (Ern´ ee and Langov´ a 2020) may have directly benefited its population, resulting in an abundance not only of traded goods, but also of other resources, including food. The overall higher amount of animal protein in the diet of the Mikulovice population compared to the Z´ alezlice LBA site could thus be a result of this favourable location. In Moravia, the context of LBA burials is less clear, as the interpretations of the Bluˇ cina Cezavy finds range from ritual sacrifice to war event (Salaˇ s 2023). In any case, compared to the EBA sample, they do not seem to be as deprived of animal products as their Bohemian counterparts from Z´ alezlice. This may suggest that we are either dealing with a much more complete profile of the LBA population, or with a specific group of people somehow associated with the hillfort. Fig. 5. Carbon and nitrogen isotopic data from particular chronogeographical contexts; A: EBA Bohemian site of Mikulovice; B: EBA Moravian site of Tuˇ rany with respect to the age and the type of burial; C: LBA Bohemian site of Z´ alezlice with respect to the type of burial; D: LBA Moravian sites. S. Drtikolov´ a Kaupov´ a et al. Journal of Archaeological Science: Reports 66 (2025) 105335 9