Materials

Sample selection.

This case study focuses on personal, decorative, and other small copper-alloy objects from La Tène cemeteries from the Czech Republic, dated to the 4^th and 3^rd centuries BCE (Fig 2). The selection of 399 samples (S1 Table) was based on the chronological determination, social context, and relevance to investigated technological processes. The assemblage included 153 samples from Moravian cemeteries, 177 samples from similar sites in Bohemia, and 69 samples from the agglomeration of Němčice nad Hanou (Table 1) which included also 30 samples of imported copper-alloy coins from the Mediterranean area, from the collections of the Moravian Museum in Brno, Czech Republic. No permits were required for the described study, which complied with all relevant regulations. All samples were archaeologically contextualised in terms of their location, typology, and chronological classification.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Sample distribution by century (4^th, 3^rd, 3^rd–2^nd BCE), region (Bohemia/Moravia), and site type. The ID numbers refer to numbers in Fig 1.

https://doi.org/10.1371/journal.pone.0352885.t001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. La Tène copper alloys from the 4^th - 2^nd century BCE from the study area.

1-11 - 4^th century BCE, 12-18 - 3^rd century BCE (12-15 Plastic-Style jewellery), 20-22 – agglomeration of Němčice, 19, 23 – Mediterranean coins from the agglomeration (photo: D. Bursák, graphics: A. Danielisová).

https://doi.org/10.1371/journal.pone.0352885.g002

Dating of samples within the Central European La Tène period relative chronology [22] was used as follows:

• 4^th century BCE (410 - 300/290):
  • LT B1a (410–390 BCE)
  • LT B1 (400–350 BCE)
  • LT B2a (350–300 BCE)
• 3^rd century BCE (300/290–200/180 BCE):
  • LT B2b (300–260 BCE)
  • LTB2b-C1 (260–230 BCE)
  • LT C1 (230–200/180 BCE)
• 2^nd century BCE (200/180–120 BCE)
  • LT C2 (200/180–120 BCE)

Methods

To ensure the integrity of the samples, material was extracted from the metal core, avoiding corrosion layers [23]. Between 30 and 50 mg of alloy was collected for analysis, following standard ICP-MS protocols, including dissolution in HNO3 and Pb separation. Analytical methods included NEPTUNE Plus MC-ICP-MS for lead isotope analysis and ICP-MS for chemical composition studies, to investigate the artefacts’ composition and potential provenance. A smaller subset of samples was tested using both ICP-MS and NAA.

The isotopic and element concentration analyses were performed using Neptune MC-ICP-MS (Thermo-Fisher Scientific). The external precision for the MC-ICP-MS measurements of ²⁰⁶Pb/²⁰⁴Pb and ²⁰⁷Pb/²⁰⁴Pb ratios are below 0.006 (2SD) and of ²⁰⁸Pb/²⁰⁴Pb below 0.01 (2SD), respectively, as determined from long-term reproducibility of SRM NIST 981.

Data evaluation

Chemical concentrations

The results of the elemental analysis of the samples are presented in S1 Table. A distinct trend in the proportions of the main alloying elements (copper, tin, and lead) is observed in the gradual incorporation of lead into copper-tin alloys—a practice seen only in outlying samples in the later part of the 4^th century BCE, but becoming more systematic in the 3^rd century BCE [5]. The emergence of leaded bronze technology appears to have become more widespread by the very beginning of the 3^rd century BCE in Moravia (Fig 3.A), where a progressive increase in lead content in the alloy is evident over time. In contrast, the adoption of leaded bronze technology in Bohemia seems to have lagged behind Moravia. Highly leaded objects found in Bohemian graves were more likely imported rather than indicative of local technological expertise. However, by the second third of the 3^rd century BCE, the practice of casting bronze jewellery with added lead had become a well-established technological process within the study area.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Evolution of the lead content in the bronze alloy of the personal objects in burials through the 4^th and 3^rd centuries BCE in Bohemia and Moravia (A; filled plots and symbols - Bohemia; open plots and symbols - Moravia), in the agglomeration of Němčice (B-C); and in the imported Mediterranean coins (C).

https://doi.org/10.1371/journal.pone.0352885.g003

The lead content of the alloys from the agglomeration at Němčice nad Hanou (Fig 3.B) was generally observed to be higher than that of objects from the cemeteries. However, no chronological trends were identified throughout the 3^rd century BCE, including the 2^nd-century BCE phase, which is not contemporary with the cemeteries. The lead content of Mediterranean coins (Fig 3.C) varies: it is negligible in Ptolemaic coins, comparable to the production at the agglomeration in Syracusan coins, and significantly higher in Carthaginian coins, where the lead content regularly exceeds 40 wt%.

In terms of classifying trace element contents based on the characteristics of potential ores, the data were assembled in several correlation matrices (Tables S1–S4 in S1 Text) and subsequently visualised in a ternary system consisting of Co, and Ni, typical for chalcopyrite ores, Ag, As, and Sb for fahlore-type ores, and Zn for galena ores (Fig 4). It is evident that the profile of the production in the 4^th century BCE is more heterogeneous, indicating diverse mineral sourcing, compared to the more systematic and less heterogeneous profile of the alloys in the 3^rd century BCE. When examining individual sites within this ternary system (Figs. 50 and 51 in S1 Text), several specific patterns emerge, such as the homogeneity of assemblages from Vliněves site in the 4^th century BCE or Nechvalín site in the 3^rd century BCE. Conversely, sites like Brno-Maloměřice, Holubice and Mlčechvosty in the 4^th century BCE, and Holubice, Pustiměřské Prusy, Mistřín and to some extent Němčice nad Hanou agglomeration in the 3^rd century BCE, tend to exhibit significant non-specificity (i.e., they are not distinctly characteristic for any mineral ore type within the ternary system).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Ternary diagrams showing the relative proportions of selected trace-element groups in copper-alloy artefacts from the 4^th century BCE, the 3^rd century BCE, and the 3^rd–2^nd century BCE.

The ternary coordinates are defined as Σ(Ag + As + Sb) (wt.%), Σ(Co + Ni) (wt.%), and Zn (wt.%). Colours indicate region and symbols indicate site category.

https://doi.org/10.1371/journal.pone.0352885.g004

The PCA analysis was run on five data matrices: a) cemeteries in the 4^th and the 3^rd century BC from Bohemia and Moravia, b) cemeteries in the 4^th century BCE from Bohemia and Moravia, c) cemeteries in the 3^rd century BCE from Bohemia and Moravia, d) Moravian cemeteries and the agglomeration of Němčice nad Hanou in the 3^rd century BCE, and e) Němčice nad Hanou agglomeration and Mediterranean coins in the 3^rd and the 2^nd centuries BCE. Presented here are results of runs performed on selected log-transformed trace elements. Detailed protocols on all PCA runs are provided in S1 Text.

1. a) Analysis of the whole dataset brought forth the general information about the character of copper alloys and their differences between chronological period and geographical regions. The intermediate PC1 and PC2 values show diverse trace element profiles. Overall, the pattern points to the differences between Bohemia and Moravia (Fig 5.A), the latter showing a tendency toward a more clustered pattern. The clustering pattern is more evident also towards the 3^rd century BCE (Fig 5.B), especially in phases LT B2b-C1 and LT C1 (Fig 5.C). Convergent and divergent trends among individual sites from both regions are shown in separate site-based PCA Supplementary Fig. 52 and Fig 53 in S1 Text; this tendency becomes more apparent when smaller and more focused datasets were tested. Overall, the dispersion as opposed to shrinkage of samples in the PCA space suggest likely blended or regionally variable input in earlier phases, pointing to widely distributed production networks or non-specialised trade exchange developing into evidence of more homogeneous production towards the later phases.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Results of the PCA analysis of the complete dataset showing basic data patterning.

A – between the two main regions, B – between the two main periods, C – selected chronological phases. Site-based PCA plots for the complete dataset are provided as separate figures in Supplementary Information 2.

https://doi.org/10.1371/journal.pone.0352885.g005

1. b) The pattern for the 4^th century BCE in both regions shows the moderate data patterning throughout (Fig 6), possibly representing transitional, mixed or non-distinctive sources (Fig 6.B). The most pronounced differences are visible at the regional level: Moravian samples are virtually absent from the upper-right quadrant of the plot (Fig 6.A), while site-level variability is shown separately in Supplementary Fig. 52 in S1 Text. It is apparent the common features of copper alloy recipes between the large sites, such as Prosmyky and Brno-Maloměřice compared to comparatively smaller communities such as cemeteries in central Bohemia, while Moravian smaller cemeteries have the tendency of clustering along the main distribution of Brno-Maloměrice. Several samples from Prosmyky in the similar space as occupied by other Bohemian cemeteries show likely mixing of different production circles. Overal, however, the production in the earlier phase shows general unspecificity pointing to local procurement and manufacturing strategies characteristic for individual local regions that tentatively interconnect several sites together. The two larger communities, Prosmyky and Brno-Maloměřice, and their overlap possibly shows the wide distribution of supplying networks and high degree of interconnectedness visible at the scale of larger communities.
2. c) In the 3^rd century BCE, the difference between the two regions becomes most apparent. There are comparatively less samples from Bohemia, but it is apparent that the production remains in the traditional line of sources characteristic by Ag, Zn, and Bi, while in Moravia, the previous tendency towards high PCA scores for Ni, Sb, Co, and As remains similar (Fig 7.A). The clustering of Moravian sites within the moderate PCA results is lead by later sites, such as Mistřín, Hustopeče, and Brno-Maloměřice, but a very strong overlap is still apparent between the Moravian sites, specifically Nechvalín, and the site of Prosmyky in Bohemia, in that period represented only by several graves.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Results of the PCA analysis of the cemeteries in the 4^th century BCE in Bohemia and Moravia.

A – pattern for the two main regions. B – selected chronological phases (LT B1a, LT B1, LT B2a). Site-based PCA plots for the 4^th-century BCE cemetery assemblages are provided separately in Supplementary Fig. 52 in S1 Text.

https://doi.org/10.1371/journal.pone.0352885.g006

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Results of the PCA analysis of the cemeteries in the 3^rd century BCE in Bohemia and Moravia.

A – pattern for the two main regions. B – selected chronological phases (LT B2b, LT B2b-C1, LT C1). Site-based PCA plots for the 3^rd-century BCE cemetery assemblages are provided separately in Supplementary Fig. 53 in S1 Text.

https://doi.org/10.1371/journal.pone.0352885.g007

Divergent tendencies of individual sites, such as Ponětovice, Pustiměřské Prusy, Holubice, and the Bohemian sites of Jinonice and Vliněves, are shown separately in Supplementary Fig. 53 in S1 Text. As for dating, the PCA distribution appears to contract in the later phases represented in the figure, especially LT B2b-C1 and LT C1 (Fig 7.B).

1. d) Groupings performed for burials and the agglomeration in the 3^rd century BCE in Moravia identified similar significant trace elements and their correlations (Fig 8). Most of the variability along one axis is explained by the negative correlations of Ag, As, Sb, and Bi with the positive correlations of Zn and Ni. The second axis is characterised by the correlation of Co and Fe, although this relationship is less pronounced. The data are visually separated into two groups, a distinction further validated by K-means clustering (Fig 8.A), which identified two clusters as the optimal number. Within these two clusters, several patterns emerged: i) samples from the agglomeration were exclusively concentrated in cluster 2; ii) cluster 2 encompassed all dating phases, including the latest production phase occurring solely at the agglomeration, whereas cluster 1 intriguingly included only graves from the LT B2b and LT C1 phases, omitting the intermediate LT B2b-C1 phase (Fig 8.B); iii) site-based patterning is shown separately in Figs. 37–42 in S1 Text; and iv) when the resolution was refined to the level of individual graves, a clearer pattern emerged: samples in cluster 1 came from various sites (Pustiměřské Prusy, Mistřín, Brno-Maloměřice, Ponětovice, Hustopeče, Němčice nad Hanou burials and Holubice), whereas cluster 2 predominantly included samples from the same cemeteries—particularly those where sampling from multiple graves was possible (Figs. 37–42 in S1 Text).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Results of the PCA and K-means clustering analyses for the cemeteries and the agglomeration of Němčice nad Hanou in the 3^rd–2^nd centuries BCE.

A – K-means clustering result based on PCA factor scores. B – PCA results for selected chronological phases (LT B2b, LT B2b-C1, LT C1, LT B2-C2, LT C2). Site-based and grave-level patterning is provided separately in Figs. 37–42 in S1 Text.

https://doi.org/10.1371/journal.pone.0352885.g008

1. e) The results for the analysis of the agglomeration and the Mediterranean coins indicate that the copper alloy production at the agglomeration and the imported coins did not share the same material characteristics (Fig 9). While trace elements such as As and Co are associated with the imported coins, the alloys produced at the agglomeration are characterised by Ni, Sb, Ag, and Bi. These elements are typical of the La Tène period copper industry and suggest that the Němčice agglomeration likely utilised the same copper sources as most workshops producing jewellery found in cemeteries. Some Němčice products overlap with the distribution of values for the Mediterranean coins, but a clear distinction exists between the two production circles.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. Results of the PCA analysis for the agglomeration of Němčice nad Hanou and Mediterranean coins.

https://doi.org/10.1371/journal.pone.0352885.g009

Lead Isotopes

The lead isotope (LI) systematics of studied artefacts is provided in the Supplementary Information (S2 Text) and can be summarised as follows. The value ranges for the isotopic ratios across all samples are ²⁰⁶Pb/²⁰⁴Pb: 18.14–18.83; ²⁰⁷Pb/²⁰⁴Pb: 15.55–15.68 and ²⁰⁸Pb/²⁰⁴Pb: 38.08–38.87. Due to the wide temporal range of the data and the focus of the study, we primarily consider the chronological and spatial context.

In the chronological phases corresponding to the 4^th century BCE (Fig 10.A), any specific patterns in data distribution are not observed; rather, a natural trend from less radiogenic to more radiogenic values was recorded. The two regions, Moravia and Bohemia, follow more or less the same pattern. The only slightly noticeable cluster of data falls within the range of approximately ²⁰⁶Pb/²⁰⁴Pb 18.5–18.7 and ²⁰⁷Pb/²⁰⁴Pb 15.66–15.68 in the more radiogenic range suggesting older deposits of Variscan age. By the 3^rd century BCE (Fig 10.B), specific trends emerge. A narrower trend line among the less radiogenic values, which become more dominant, culminating in a specific LI cluster (“Cluster” in Fig 10:B) at ²⁰⁶Pb/²⁰⁴Pb 18.4–18.55 and ²⁰⁷Pb/²⁰⁴Pb 15.61–15.63 can be observed. Above a ²⁰⁷Pb/²⁰⁴Pb value of ca 15.64, the lead isotope systematics is more or less similar to that of the 4^th century BCE (Fig 10.A). This LI cluster is to a large extent filled with samples dated to later two thirds of the 3^rd century BCE.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 10. Lead isotope (²⁰⁷Pb/²⁰⁴Pb vs. ²⁰⁶Pb/²⁰⁴Pb) systematics of the studied artefacts.

A–B: Lead isotope ratios for the 4^th (A) and 3^rd–2^nd (B) centuries BCE. C–D: Same plots, with colour scale indicating lead content in wt.% of the analysed alloys (<5%, > 15%, intermediate values in between). Symbol shapes distinguish site categories (cemeteries, Němčice, and Mediterranean coins from Němčice). Purple-highlighted fields mark LI cluster discussed in the Results section.

https://doi.org/10.1371/journal.pone.0352885.g010

It is evident that in the 3^rd century BCE, the lead isotope systematics is significantly influenced by the lead content in the artefacts, markedly exceeding the threshold of 5 wt.% (Fig 10.C-D). Beyond this level, the alloy was deliberately leaded, and the lead isotopic signatures thus correspond to lead-bearing ores. Artefacts outside the LI cluster and a trend line of the 3^rd century BCE only rarely exhibit lead contents exceeding 5 wt. %. Notable exceptions are Mediterranean coins, particularly those originating from Carthage or Syracuse (Fig 3.C).

Across sites, the LI ratios organise into specific groups, pointing to differences in material provenance and source homogeneity. In the 4^th century BCE (Fig. 8–10 in S2 Text), large datasets such as Brno-Maloměřice, Prosmyky, and Jenišův Újezd exhibit significant heterogeneity, possibly caused by a mixed supply of raw materials. In the following century (Fig. 8, 11, 12 in S2 Text), these large datasets display much tighter clustering within the field of the LI cluster. Interestingly, these larger sites gradually exhibit greater homogeneity compared to smaller datasets from sites like Hustopeče, Holubice, Vliněves and Prague. The LI cluster appears to be more typical for Moravian sites, while the only Bohemian cemetery sharing this pattern is Prosmyky. In this respect, Moravia appears to be more isotopically unified than Bohemia.

Another phenomenon observed in the 3^rd century BCE is that only the datasets from Němčice nad Hanou agglomeration and Mistřín cemetery substantially exceed the distinguishing ²⁰⁶Pb/²⁰⁴Pb value of 18.55—a threshold that defines the LI cluster. These sites, along with a few samples from Holubice and Blučina, appear notably more heterogeneous in terms of lead isotope composition (Fig. 8, 11, 12 in S2 Text). In Němčice nad Hanou agglomeration, the earliest LT B2b–C1 finds fall outside of LI cluster, whereas the later LT C1 artefacts align with this isotopic field (Fig 11).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 11. Dynamics of technological innovations at Moravian cemeteries and the agglomeration of Němčice nad Hanou.

It is apparent that the agglomeration largely adopted a new source of lead (“LI cluster”) only in the second half of the 3^rd century BCE.

https://doi.org/10.1371/journal.pone.0352885.g011

Using the provenance model described in the methods section, following results were obtained, where individual datasets exceeded the set probability threshold of 10% for all three lead isotope ratios with ²⁰⁴Pb (Fig 12). This analysis was applied to all individually tested data subsets based on selected categories (cemeteries, the Němčice agglomeration, and Mediterranean coins, each chronologically categorised where possible). The provenance model found 87 positive matches with available ore deposits database from all 339 inspected items (ca 25% of total). For all non-Mediterranean artifacts, the highest level of consistency was observed with deposits in the Harz Mountains. Cemeteries from the 4^th century BCE exhibit higher source heterogeneity with only 38 matches with defined ore regions (17% of 4^th century BCE samples), with 34 cases linked to deposits in the Harz (18 cases), the Balkan/Rhodopes (14 cases) and Cyprus (6 cases). In the 3^rd century BCE, the situation became more focused, with 28 samples matching comparable ore datasets (c. 20% of 3^rd-century BCE samples). Almost all of these matches point to Variscan ore provinces in central Europe, with 26 cases compatible with ore fields in the Harz Mountains and only two observations consistent with Cypriot deposits, although a more detailed inspection of traditional overlap plots with individual ore regions shows that several sub-provinces of the Rhenish Massif fulfil similarly strict compatibility criteria as well (Fig. 19a and 19b in S2 Text). The narrower geochemical definition of the 3^rd-century BCE assemblages is also supported by the trace-element analysis and by the internal structure of the data. For the Němčice agglomeration, the 3^rd-century (including the 3^rd–2^nd-century) material already shows a mixed but clearly Harz-dominated provenance: seven samples match Harz-type ore fields and three fall within Rhodope-type fields. The 2^nd-century BCE phase adds one further match to each of these provinces, so that both Harz and Rhodope-related signatures are present throughout the later history of the site. However, phase-wise, detailed analysis of lead-isotope systematics from Němčice agglomeration reveals a more complex, internally differentiated artefact provenance with distinct trends. These fine-scale patterns are not captured at this level by the provenance model and are discussed in detail in the Supplementary Information (Fig. 13 and 14 in S2 Text). Rhodope-type signatures still form the single largest group (6 of 12 coins), but the updated database also points to additional Mediterranean sources in the Betic Cordillera (5 cases) and the Apennines (1 case). Overall, it is evident that in the 3^rd century BCE, it is possible to provenance more objects than in the preceding century. These results suggest that during the 4^th century BCE, metal resources exhibit a greater tendency toward source mixing. In contrast, production in the 3^rd century BCE reflects a more streamlined process, more likely utilising primary sources from the Rhenohercynian Fold Belt. It is evident that a significant share of the data, specifically 75%, remains without a clear provenance.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 12. KDE based provenance analysis results with respect of the Pb amount in alloy.

Green: Pb < 5 wt. %.; orange: Pb > 5 wt. %.

https://doi.org/10.1371/journal.pone.0352885.g012

Topography of technological innovations

Our results reveal a pronounced diachronic shift in the production of copper‑alloy artefacts between the 4^th and 3^rd centuries BCE across the regions studied in Central Europe. Assemblages dated to the 4^th century BCE exhibit wide dispersion in both lead isotopes and trace-element space. Regionally, Moravia tracks more clearly toward compositional structure in the 3^rd century BCE than Bohemia, perhaps reflecting earlier or stronger engagement with transforming networks that re-organised alloying and procurement of raw materials.

At the agglomeration of Němčice, early artefacts (phases LTB2b, LTB2b-C1) bear more radiogenic signatures aligned with the Alpine/Carpathian field; later LT C1 items align with the 3^rd-century LI cluster observed at cemeteries, reflecting a transition from distinct, possibly mixed supply arrangements to the broader regional consolidation centred on a dominant lead source. The sequence is consistent with the defining third-century lead signature appearing first outside the agglomeration (Fig 11).

This diachronic consolidation is evident not only across regions but also within individual cemeteries (Figs 8, 11). Additional patterning emerges when the PCA and K-means clustering results are mapped geographically (Fig 13). Sites with a chemical composition distinct from the agglomeration at Němčice nad Hanou (K-means cluster 1) are predominantly located in the northern part of the study area, along the presumed south-north communication route. Furthermore, the earliest examples of this category (Němčice nad Hanou burials, Pustiměřské Prusy, Holubice, dating to the early 3^rd century BCE) are situated in the northeastern portion of this distribution, while the most recent burials (Brno-Maloměřice and Ponětovice, from the second half of the 3^rd century BCE) are found in the western part of the study area. In contrast, cemeteries containing materials similar to the later production at the Němčice agglomeration (K-means cluster 2) are mainly concentrated in the South Moravian lowlands and the southern part of the study area, although individual burials extend northward. One notable exception to the otherwise distinct geographical separation of clusters is Brno-Maloměřice (and to a little extent also Ponětovice), which features graves from both clusters and all chronological phases, which is possibly caused by the character and the size of the site. It means that communities represented by K-means cluster 2 were likely part of a more centralised manufacturing system, sharing access to common raw materials and technologies, with a distribution focused in the South Moravian lowlands. In contrast, K-means cluster 1 may reflect a more decentralised mode of production or reliance on different sources, evident in its spread along the presumed south–north communication axis. The presence of both clusters across all phases at Brno-Maloměřice suggests it served as a contact zone or transit hub where distinct traditions intersected and cultural flexibility was expressed. Already in the LT B2b phase, several sites show assemblages aligned with new technological trends, whereas broader production still combines these innovations with traditional lead sources. Within Bohemia, burials at the cemetery of Prosmyky is the clearest regional, albeit isolated, expression of these progressive trends (Fig. 8, 11, 12 in S2 Text). By the later 3^rd century BCE, communities represented in the cemeteries had largely abandoned the traditional sources and shifted to one dominant lead source (Figs 11-13). These differences could reflect the variable timing and degree of integration into shared supply networks, the weight of local recycling, or differential mobility regimes. Importantly, such heterogeneity/homogeneity is mirrored across isotopes and trace elements, strengthening the inference that genuine differences in access and practice rather than statistical noise are being observed.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 13. Regional patterns in the emergence of different alloy recipes, reflecting the adoption of new technological trends by cemetery communities (map created in ArcGIS Pro, background: DMR 5G Map by ČUZK, 2026, www.cuzk.gov.cz, CC-BY 4.0; graphics by A.D.).

https://doi.org/10.1371/journal.pone.0352885.g013

Production at the agglomeration adopted this major source as well, although it continued in parallel to draw on lead from a more radiogenic isotopic zone, largely abandoned by cemeteries by this time (Fig 11), likely indicating continued mixed inputs or stronger effects of recycling and local mediation. As mentioned above, in the earliest phase at Němčice nad Hanou agglomeration, the adoption of leaded alloys is not apparent; rather, leaded bronze production expanded later as metallurgical production scaled up to meet the increased demand. This timing is evidence of an emerging central-place role consolidating throughput—storage, brokerage, and high-volume casting—after the shift was already underway in surrounding communities. The results challenge a simple model in which agglomeration sites originate technological change later seen in rural communities. Instead, funerary contexts, representing broader community practices, register the transition from flexible, mixed 4^th-century production to more structured 3^rd-century alloying, with agglomerations subsequently amplifying and scaling these shifting regimes. It should be noted that the production spectra of the cemetery communities and the agglomeration largely overlapped. In both contexts, copper-alloy metalworking was primarily directed towards the manufacture of jewellery, personal ornaments, and decorative objects more broadly. The earlier phases of production at the agglomeration provide evidence for the casting of elaborate costume components, including decorative chain belts, which are also represented in contemporary burials. Thus, although differences between funerary and settlement contexts may complicate direct comparison of the technologies employed, the same categories of objects occur in both environments. Even in cases where contemporaneous cemetery jewellery does not fully mirror agglomeration outputs, the results imply partially overlapping but distinct workshop networks and procurement channels. Glass production and coinage, hallmarks of later agglomeration specialisation [6] postdate the alloying transition documented in cemeteries already about one or two generations earlier.

The disparities between Bohemia and Moravia are best explained by their geographic positions and shifts in the orientation and organisation of local networks. During this period, the Amber Road and Middle Danube corridors rose in importance, while Bohemia—lying farther from these routes—functioned largely as a transitory zone linking distant raw-material sources to the west. Only larger communities organised around centralised hubs, such as Prosmyky, were able to track and adopt emerging innovations. The same socio-economic pattern appears in other domains, notably diet: gains in agricultural efficiency were concentrated in the Middle Danube area and were taken up only by selected large communities in Bohemia [34].

Provenance of the raw materials

Provenance analysis with very large comparative datasets risks uneven spatial representation of ore provinces and an illusion of objectivity if matches are taken at face value. Three confounding processes merit consideration. First, recycling can re-average signatures; however, the tight 3^rd-century clustering across both isotopes and trace elements is difficult to produce through recycling alone and instead suggests consistent inputs and recipes. Second, mobility might seed non-local signals in otherwise local assemblages; yet the systematic nature of 3^rd-century convergence (and the weak Mediterranean copper-alloy signal) implies that long-distance movements did not dominate the procurement underpinning mass leaded bronze production. Third, sampling intensity varies by site; therefore, the aim in this study was to refrain from over-interpreting single outliers and to emphasise patterns replicated across multiple cemeteries and contexts. The approach presented in this study mitigated these issues by combining a KDE match/no-match framework with traditional bivariate distribution analysis and by cross-checking isotopic inferences against trace-element behaviour and archaeological context [1,5,20]. This triangulation increases confidence in (a) a multi-source, flexible 4^th-century regime and (b) a 3^rd-century consolidation around Rhenohercynian ore provinces for the lead component—while acknowledging inevitable blurring from recycling and mixing. However, it is essential to take into account that the model does not provide a single definitive explanation of provenance but primarily serves as a tool for working with the vast comparative database available today.

The comparisons of assemblages against geological fields show two principal isotopic “domains” (Fig. 15a–17c in S2 Text). Less radiogenic signatures (≈²⁰⁶Pb/²⁰⁴Pb < 18.55) overlap with nearby Transalpine provinces (Bohemian Massif, Rhenish Massif, Harz Mountains) and typically present lower As–Sb–Ag with a mild propensity for Zn. More radiogenic signatures (≈²⁰⁶Pb/²⁰⁴Pb > 18.55) overlap with younger Alpine/Carpathian belts (Balkan/Rhodopes; Valais; Apuseni/Slovakia) and often show higher Ag and Sb, consistent with a fahlore-type component. These relationships parallel previously observed clusters in the 4^th century BCE Duchcov hoard (Cluster 3 ≈ Bohemian Massif/Erzgebirge; Cluster 1 ≈ western Alps) [1], suggesting that the earlier craftsmen drew on multiple ore inputs, signalling mixed resources, flexible alloying, and a lack of consistent supplies or recipes—conditions consistent with small-scale and locally mediated production, aligning with the prevailing (often implicit) model of Iron Age organisation of the production. The emergent pattern in the 3^rd century BCE, especially observed in Moravian cemeteries and the Prosmyky site in Bohemia, includes a characteristic reliance on a major consistent lead source (LI cluster). In geological terms, the source regions of the LI cluster can be situated within the Rhenohercynian Fold Belt, in particular parts of the Rhenish Massif and the Harz Mountains. This Variscan fold-and-thrust belt is dominated by Devonian–Carboniferous sedimentary successions, within which late Variscan to Permian hydrothermal circulation produced significant Pb–Zn–Ag vein and stratiform mineralisation (regional overviews: [35,36]). A detailed comparison of artefact Pb isotope compositions with these ore fields indicates the closest matches with hydrothermal vein systems in the Taunus segment of the Rhenish Massif (recorded as “Hesse” in our dataset) and with deposits in the Harz (Fig. 19a and 19b in S2 Text) [35,37–42]. Previous Pb isotope analyses of rouelles (spoked wheel amulets), beads and other La Tène-period lead objects from the Wallendorf site show that their compositions fall within the Variscan and post-Variscan deposits of the Rhenish Massif, leading to the assumption that Roman exploitation of Eifel lead largely continued an already established Celtic mining tradition [35]. Taken together with results in this study, these data are consistent with a long-lived lead-supply system centred on these polymetallic deposits, plausibly operating from at least the 3^rd century BCE and persisting into the 1^st century BCE and the early Roman period, thereby extending and strengthening the hypothesis of pre-Roman Pb procurement. At present, there is still no direct evidence for Iron Age extraction of the Pb–Zn–Cu sources in the region of Rhenish Massif and Harz. Nevertheless, iron-smelting furnaces are documented in the neighbouring Siegerland region for La Tène B2 (middle of the 3^rd century BCE), together with tentative indications of copper smelting [43]. Further evidence for intensive Iron Age exploitation in the Rhenish Massif is linked to the salt economy: at the eastern margin of the Taunus, the large salt-evaporation complex at Bad Nauheim documents sustained salt production from at least the 3^rd century BCE, i.e., contemporaneous with the lead procurement regime considered here [44]. If we reconsider the potential long-distance connections linking these mining and salt-producing zones to Bohemia, salt again offers an obvious candidate vector: earlier work on La Tène communication routes has envisaged salt being brought to Bohemia from Saxony, usually hypothetically located in the Halle/eastern Harz region [45,46], even though explicit archaeological evidence for salt production in central Germany between the 4^th and 1^st centuries BCE is largely lacking [47]. In light of the new isotopic and archaeological evidence presented here, it is suggested that, in addition to the salt transportation, the same corridors may also have channelled Pb-bearing materials derived primarily from Rhenish Massif deposits, providing a new hypothetical linkage between Bohemia and the Rhenish Massif (with a possible, though less likely, contribution from the eastern Harz region). If we complement the archaeometric signal with the currently documented pattern of extractive and industrial activity, the balance of probability between these alternatives favours the above discussed deposits in the Rhenish Massif rather than the Harz as the principal Pb source.

Despite increased contact with the Hellenistic world, copper-alloy data show only sparse matches to Greek/Macedonian provinces, and those are mostly confined to 4^th-century items (e.g., few Balkan/Rhodopes matches). In geological terms, the “Rhodope” matches identified by the provenance model point to the Tertiary polymetallic Pb–Zn–Ag (±Cu ± Au) ore systems of the Thrace–Rhodope belt—hydrothermal vein, replacement and epithermal deposits hosted in the metamorphic core complex of the Rhodope and Serbomacedonian massifs [27,28,48–51]. Mediterranean coins display isotopic signals distinct from southern copper-alloy production, however, there is no evidence of them being used as a material source intended for recycling (Fig 9). In other words, the widespread 3^rd-century adoption of leaded bronze in Central Europe is not anchored in Mediterranean metal flows; rather, Transalpine ore fields, especially the Rhenohercynian Fold Belt, dominate the provenance of the lead component that enabled the increased production of the Plastic-Style jewellery. This observation matters for broader debates on technological transmission. If the mechanism had been import-driven through Mediterranean artefacts, earlier objects would prominently carry Mediterranean signatures. Instead, present dataset favours a model in which stylistic ideas and certain technical concepts travelled widely, but the material substrate (ores) for Central European mass production derived predominantly from Transalpine sources.

Socio-economic and cultural implications of emergent leaded bronze production

Our results reveal that the selection of leaded alloys was a deliberate and selective process, employed when the jewellery shape and decorative requirements necessitated a material with specific properties. In contrast, more common items, such as brooches, do not show evidence of leaded alloys. This indicates that bronze workshops had mastered multiple techniques and that craftsmen made intentional material choices based on the functional and aesthetic demands of the fabrication process. Consolidation around few lead sources and alloy recipes implies organised access and control of supply/distribution for specific markets, connected with the onset of the popularity of Plastic-Style jewellery. Lead’s lower price relative to tin [14] reduced unit costs for large, voluminous ornaments, enabling a surge in robust jewellery forms in the course of the 3^rd century BCE. The result is a compelling fit between alloy properties, cost structure, and stylistic demand. From the visual perspective, the new, attractive jewellery made in the Plastic-Style was significantly more robust and likely functioned as an important status symbol of the new, economically and perhaps even politically more established and self-confident Iron Age society of the 3^rd century BCE. The idea of a new artistic style and crafting probably originated in the Mediterranean, while the material used for these artistic manifestations had a local origin, independent of the Mediterranean world. Gradually, with the growing demand for new jewellery designs made from new materials, this development became fertile ground for social stimulation. The demand for new materials largely increased in volume which in turn created an opportunity to be seized by individuals or groups connected to long-distance socio-economic networks, possibly fostered by previous military campaigns. Archaeological evidence points to local warrior elites, possibly the heirs of 4^th-century BCE non-centralised, rural based society, as the most likely enforcers of the new economic opportunities that has given rise to influx of new, much more conspicuous bronze ornaments.

We interpret the new leaded bronze jewellery (and, shortly thereafter, also newly introduced glass, and re-introduced amber ornaments) as archetypal “bulk luxury commodities” that encode prestige through form and style, yet are sufficiently affordable, thanks to standardised inputs and scalable craft, to diffuse broadly [52–54]. Archaeometrically, 4^th-century bronzes read as heterogeneous, locally mediated, and weakly coordinated utilitarian goods; by contrast, 3^rd-century leaded bronzes combine (i) prestige coding (Plastic-Style and its characteristic bulkiness), (ii) concentrated sourcing (narrow Pb isotope field, and similar alloy recipes), and (iii) standardised distribution enabled by cost advantages—all the hallmarks of bulk luxury goods.

Several comparative analogies strengthen this reading. First, Bohemian sapropelite jewellery [55] represented a locally abundant material entering wide networks via distinctive, fashionable forms, similarly fostering elite brokerage over the flow and distribution rather than exclusive control of rare inputs. Together with early leaded bronze ornaments, it is one of the earliest archaeologically visible indicators of this process, rooted in locally or regionally available, non-Mediterranean resources. Second, the late 3^rd century BCE spread of glass ornaments expresses analogous standardisation and regional integration of a commodity with both exotic connotations and wide availability [56]. Rolland’s analysis of Late Iron Age glassmaking shows that production was closely tied to rising demand from newly emerging economic elites – the homines novi of the later 2^nd century BCE – highlighting the transformative power of bulk-luxury goods [57].Third, although amber is not systematically attested within the cemetery communities examined here, it appears to have gained renewed popularity through central places, as recently confirmed by evidence for amber-working at agglomerations such as Němčice nad Hanou [10]. Amber therefore represents a useful example of a bulk-luxury commodity curated through organised networks of supply, production, and distribution. Raw amber was imported from northern source regions and transformed into desirable finished products in workshops located at Middle Danube agglomerations. This pattern was later continued and further developed within the trade and production systems of the oppida [58]. The renewed prominence of amber may thus reflect an intensification of connections along the north–south axis, represented by the Amber Road, which complemented the already developed west–east routes used for the movement of metals, salt, and other commodities. Finally, coinage made of precious metals is another example of a comparably standardised system that emerged during the 3^rd century BCE, reflecting broader shifts toward technological cohesion, material integration, and economic coordination, however based on different networks of sourcing and production [6]. These bulk-luxury goods formed an intermediary category: costly enough to mark prestige yet widely accessible, combining aesthetic capital with broad reach and thus enabling new modes of distribution, regional market consolidation, and the emergence of new elite configurations.

Central places, including Němčice nad Hanou, are best viewed as throughput hubs within an already changing landscape, scaling up production, buffering supply fluctuations, and exporting recognisable stylistic packages across regional networks. Operating as gateway nodes on long-distance routes, these sites used control over raw-material access, transit points, storage, and brokerage to turn flows of leaded bronze and, later, glass, amber and precious metals for coinage into stable rents and market power. As prestige-coded but widely accessible goods scaled up, the focus of elites shifted from monopolising rare materials to coordinating flows, crafting, and distribution at volume. This elite input in consolidating the production and distribution flows of bulk luxury goods was tightly connected with social cohesion, cultural identity, and symbolic expression, reconfiguring thus how we understand Iron Age markets and, in a broader sense, geopolitical integration of the whole regions. Crucially, at the onset of this process, when leaded bronze first entered the markets, agglomerations had not yet formed. Instead, relatively independent communities, represented by their cemeteries, secured high-volume, provenance-specific goods, while the later rise of agglomerations appears to have increased the efficiency and liquidity of these flows. Communities around cemeteries were not passive repositories of elite fashion; rather, they registered early, community-wide shifts in alloying and sourcing that preceded – and then fed into – agglomeration-based scaling. This perspective helps explain why cemetery assemblages partially overlapped with agglomeration inventories: their workshops were connected but not identical, and their procurement was integrated yet diversified. In turn, the cemetery record captures early adoption among households and kin groups, while central places optimised the production only about a generation later. This model aligns with later developments in the 2^nd and 1^st centuries BCE when late La Tène and early Germanic polities, based again chiefly in Bohemia, capitalised on extensive external supplies, as evidenced by connections to material flows from Hispania [59,60], or later making use of Roman-supplied raw materials to produce their own prestige goods [21]. This highlights durable Central European capacity to translate external connections into standardised, high-demand commodities.