Introduction

The Late Bronze Age represents a significant point of cultural and social change in European prehistory (c. 1350–800 BCE), laying the foundations for increased complexity and the emergence of political structures in the Iron Age. The increasing homogenization of material culture and the spread of urn burials that characterize the period is generally attributed to increased mobility and connectivity (Cardarelli et al. 2020; Sørensen and Rebay-Salisbury 2023). The cemetery of Inzerdorf (coordinates: E:15°42ꞌ20ꞌꞌ, N: 48°18ꞌ40 ꞌꞌ, Fig. 1) is ideally suited to examine whether the change of burial rites at the onset of the Late Bronze Age was a local development or triggered by newcomers and understand age-related and gendered patterns of mobility and shifts in land-use during the occupation of the cemetery.

Fig. 1
figure 1

Location of the cemetery of Inzersdorf, Lower Traisen Valley, Austria. Created with ArcGIS Pro v 2.9.5 and the European basemap provided by 1:10 m Natural Earth I with Shaded Relied and Water and provided permission free by Natural Earth

Full size image

The emergence of strontium isotope analyses (87Sr/86Sr) of cremated bones as a new bioarchaeological technique allows fresh insights into traditional questions of intercommunal relations (Snoeck et al. 2022). The strontium isotope ratio (87Sr/86Sr) in a region is predominately influenced by the ratio in the underlying bedrock, which then weathers into soil where it is incorporated by plants, enters the food chain and is incorporated into human and animal skeletons in the place of calcium (Bentley 2006; Capo et al. 1998; Montgomery 2010). It has been demonstrated that cremated remains are excellently suited for strontium isotope analyses because the crystalline structure of bone changes during cremation in such a way that post-burial exchanges with the soil in which the deceased has been buried are limited and can be removed by carrying out an adequate pre-treatment. (Snoeck et al. 2015).

Strontium isotope ratios in the otic capsule of the petrous portion of the inner ear have been shown to correspond to early childhood signal, whereas turnover rates for long bones are slower than those of ribs, which provide the closest to an end-of-life signal (Veselka et al. 2021b). This allows a detailed reconstruction of a person’s mobility history in relation to the individual’s age (Fahy et al. 2017).

Here, strontium isotope data of 215 calcined human bone samples was interpreted in the context of a local baseline consisting of 163 modern plant samples from 55 sampled sites in the vicinity (within 10 km), to investigate mobility and landscape use on a local and inter-regional scale. This study demonstrates the interpretative potential of combining a high-resolution mapping approach with a large dataset of 87Sr/86Sr from human remains to obtain detailed insights into the spread of the Urnfield phenomenon in a small regional context as well as highlight the implications of mobility for both individuals and the community.

The cemetery of Inzersdorf in the Traisen Valley

The cemetery of Inzersdorf encompasses 273 graves and is located in the Lower Traisen Valley near the confluence of the Traisen and Danube rivers, a geographically advantageous position. Not only are the fertile lands suitable for sustainable farming, but the rivers also provide resources, a convenient East-West route along the Danube, and access to the Alpine foothills via the Traisen.

Thanks to the large-scale rescue excavations conducted in the 1980s during gravel extraction and road construction, the Lower Traisen Valley is one of the most extensively researched archaeological landscapes in Austria, with evidence of human habitation spanning all prehistoric and historical periods (Gattringer and Neugebauer 1982). Inzersdorf, with its 273 preserved graves, remains one of the largest early Urnfield Period cemeteries excavated in Austria to date, although large parts of the cemetery were destroyed during gravel extraction (Gattringer and Neugebauer 1982; Neugebauer and Gattringer 1988).

Similar to the preceding Middle Bronze Age, the oldest graves feature rectangular grave pits and stone constructions, and contain scattered burnt bones and ceramic depositions (Lochner 2012; Sørensen and Rebay 2008). A grave with two parallel body-length features is amongst the oldest, containing the scattered cremated remains of three individuals (Fritzl 2017; Lochner 2015). The majority of the graves date to the beginning of the early Urnfield Period (~ 1350 − 1100 BCE). The urn burials contain cremated grave goods including fire-affected vessels and show little differentiation. Several graves date to the late Urnfield Period (~ 1000 − 800 BCE), again mainly urn burials with few unburnt accessory vessels (Lochner 2013). Bronze grave goods include costume components such as arm rings, Noppenringe, buttons, and pins; tools are rare, while weapons are completely absent (Fritzl 2017).

The geological makeup of Lower Traisen Valley around the site is diverse (Fig. 2); as the river flows from south to north, it traverses the geologically young Alpine foothills. The geologically old Bohemian massif dominates the geology west and northwest of the site (Wessely 2006). Reaching the Danube lowlands, gravel banks have accumulated on both sides of the river. The cemetery itself is located on such a gravel bank (Fig. 2) covered by approximately 20 cm of sandy humus. In the northeast, the lowlands have accumulated various aeolian sediments. In addition, human activities during the Holocene have significantly altered the landscape of the Traisen Valley (Wiesbauer 2019).

Fig. 2
figure 2

Bedrock geological map of the Lower Traisen Valley and surrounding region (after Federal Geological Survey of Austria 1984, 1:50,000 scale). Only sampled units are indicated by colour, unsampled units indicated by cross-hatching. Plant sampling sites indicated with black dots and the location of the Inzersdorf cemetery with a cross

Full size image

Consequently, a large range of isotope ratios may be anticipated in the Traisen Valley. 87Sr/86Sr is determined by three factors of the bedrock: formation age, initial isotope ratio, and the abundance of rubidium in its minerals (Faure 1977; Faure and Mensing 2004). The bioavailability of strontium (BASr) can vary widely due to various influencing factors, such as the specific weathering process of the bedrock or climatic factors. Additionally, what is eventually taken up and incorporated by an organism depends partly on a variety of habitual and physiognomic aspects, such as for plants the root depth or for animals and humans the specific diet or age of an individual (Bentley 2006; Oelze et al. 2012; Snoeck et al. 2020). Studies have shown that drinking water contributes little to strontium incorporation and plants generally contain much more strontium than meat, which suggests that plants provide the most important strontium source (Bentley 2006; Lewis et al. 2017). Whilst discussions on the best material to use as a proxy for isoscape mapping are still ongoing (Crowley et al. 2017; Evans et al. 2010; Frei and Frei 2011; Maurer et al. 2012; Price et al. 2002), we opted to contextualize the strontium isotope data from individuals buried at Inzersdorf with a BASr map in the region based on 163 plant samples in the Traisen Valley and beyond, following Snoeck et al. (2020). Plants form the first step in the food chain, and measurement on the whole plant represents all strontium inputs including atmospheric, geological, soil and water inputs (Burger and Lichtscheidl 2019). Furthermore, plants with diverse root depths were sampled to account for differences between the topsoil and deeper layers. This approach aims to capture variations, as prehistoric populations would have consumed a variety of crops and plants, each displaying a blend of these soil layers depending on the respective conditions. Nevertheless, it is important to acknowledge that utilizing modern plant samples as proxies carries a risk, as they may have been impacted by anthropogenic factors such as movement of soils, changing climate or fertilization. However, in this study this risk has been mitigated through meticulous selection of sampling sites and the collection of a large number of samples, which help to highlight anomalous values.

Materials and sampling strategy

The 273 graves of Inzersdorf contained the remains of 251 individuals, including 41 women, 32 men, 36 unsexed adults, and 118 subadults (see supplementary material).

As parts of the burial site Inzersdorf has been destroyed, the palaeodemographic distribution was compared with demographic proxy estimators of representation (Bocquet-Appel 2002) to check whether this sample may represent a living prehistoric population. These proxies have been frequently used for prehistoric populations (e.g. Downey et al. 2014; e.g. Kohler and Reese 2014). The estimators use ratios of age groups which are known to be less sensitive by individual age-at-death-estimation errors in general, in comparison to absolute numbers. Inzersdorf seemed to have a lack of infants and adolescents, but the ratio between subadults and adult individuals is representative, being 1:1 (Angel 1969; Bocquet and Masset 1977). This points to a methodological issue due to a high degree of fragmentation and a lack of diagnostic elements in this sample. Generally speaking, the mortality distribution of Inzersdorf is similar in comparison to mortality distributions in high-mortality populations (United Nations 1955; Weiss 1973), which makes the Inzersdorf cemetery suitable for investigating differences in strontium ratios between these age groups when methodological limitations are acknowledged.

Samples analysed for strontium isotopes were taken from each individual, whenever feasible. Sampling was merely limited by an insufficient degree of calcination of the bone, which amounted to a total of 182 sampled individuals.

The sampling strategy included collecting material from three different parts of the skeleton, allowing to infer mobility occurring at various ages. Long bone diaphyses were preferentially sampled, as they are almost always present and are among the bones with the longest turnover rates, allowing to go back 10–20 years in an individual’s life history. Eighteen rib samples provide insights into roughly the last five to ten years or so of an individual’s life due to short bone turnover rates (Fahy et al. 2017). Additional 15 samples were taken from the otic capsule of all sample-able petrous bones. The otic capsules provide in-utero to early childhood signals, as they do not remodel after about two years of life (Veselka et al. 2021a, b). Other bones such as cranial bones or unidentified fragments were only used when no other materials were preserved (see supplementary information). A total of 215 human bone samples were analysed.

163 plant samples were collected from 55 sites in an area of approximately 10 km around the site. At least two sampling locations for each geological feature were selected from the geological basemap of the greater Krems area 1:50.000 (Geologische Bundesanstalt 1984). Sampling sites were selected according to the lowest observable level of anthropogenic activity within a geological unit. At each location, three samples of grass, shrub, and trees were taken which were combined into one sample for each type, to reflect differences in root depths and provide a site-specific average. Samples were collected from sites over elevations from 191 m to 476 m.

Methods

The analysis of human remains was performed following the guidelines of BABAO (Brickley and McKinley 2004; Mitchell et al. 2017). Bone fragments were individually separated into the following anatomical areas: cranium, thorax, pelvis, upper limbs (including the shoulder girdle), lower limbs and unidentifiable long bone fragments. The total bone weight and the weight of the individual fractions were recorded. If several individuals were commingled, the MNI (Minimum Number of Individuals) was estimated based on the most frequent element and further indicators, for instance of sex or age at death, which separated two individuals from each other. Sex estimation based on morphological features followed Phenice (1969) and Walker (2008), metric methods were adapted from Cavazzuti et al. (2019). Age of death was estimated based on tooth eruption (Ubelaker 1997), the iliac auricular surface (Buckberry and Chamberlain 2002; Lovejoy et al. 1985; Osborne et al. 2004), transitional analysis (Boldsen et al. 2002), and the ossification of epiphysis (Schaefer et al. 2009). Mortality rates and mortality table were calculated based on the identified age groups and the guidelines published by Acsádi and Nemeskéri (1970).

Strontium isotope analysis on cremated human remains and plants were undertaken at the Archaeology, Environmental Changes & Geo-Chemistry Research Group (AMGC) at the Vrije Universiteit Brussel (VUB) and the G-TIME laboratory of the Université Libre de Bruxelles (ULB), both located in Brussels, Belgium.

Cremated bone fragments were pre-treated in an ultrasonication bath with three rinses of milliQ water followed by 1 M acetic acid for 7 min, and then thoroughly rinsed three times with milliQ water (Snoeck et al. 2015). Samples were then dried at 50 °C and powdered using an agate mortar and pestle. 10 mg of the powdered sample was weighed into a Teflon beaker, digested with 1mL 14 M HNO3 and then evaporated to dryness. For the plant samples, 0.25 g of plant was first microwave digested in 4mL 14 M HNO3 and 0.5mL HF in a Milestone UltraWAVE, 0.5mL H2O2 was then added, and evaporated to dryness.

For both cremated remains and plant samples, strontium was extracted through column chemistry using ion exchange resin (Sr Spec, Triskem). In short, the columns and resin were rinsed 2 x with 2 M HNO3 then 2 × 7 M HNO3, samples loaded in 7 M HNO3, charged 4 × 1mL 7 M HNO3, and columns were eluted and Sr collected with 6 × 1mL 0.05 M HNO3. Measurements of 87Sr/86Sr for calcined bone were carried out on a Nu Plasma MC-ICP-MS (Multi-Collector Inductively Coupled Plasma Mass Spectrometer, Nu015 from Nu Instruments, Wrexham, UK) at the G-TIME laboratory at ULB, while the plant samples were measured on a Nu Plasma 3 (PD017 from Nu Instruments, Wrexham, UK) at AMGC, VUB. Repeated measurement of NBS987 provided an average 87Sr/86Sr of 0.710245 ± 0.000040 (2SD for 20 analyses) at ULB and 0.710246 ± 0.000020 (2SD for 37 analyses) at VUB, which is comparable with the mean of 0.710252 ± 13 (n = 88) obtained using TIMS (Thermal Ionization Mass Spectrometry) (Weis et al. 2006). Repeated measurement of SRM1515 (Apple leaf) returned an average value of 0.713951 ± 0.000024 (2SD for 13 analyses) which is comparable to the reported value of 0.713950 (n = 24) reported by Oeser & von Blanckenburg (Oeser and Blanckenburg 2020). All the data were corrected for mass fractionation by internal normalization to 86Sr/88Sr = 0.1194. The raw data were normalized using a standard-sample bracketing method with the recommended value of 87Sr/86Sr = 0.710248 (Weis et al. 2006). For each sample, the 87Sr/86Sr value is reported with a 2SE error (absolute error value of the individual sample analysis).

Spatial analysis and predictive mapping were undertaken in ESRI ArcGIS Pro 2.8.0 using the Spatial Analyst and Geostatistical Analyst extensions. To identify geological strontium contributions, plant 87Sr/86Sr sampling sites were spatially joined to the surface geology of the greater Krems area (1:50.000), published by the Federal Geological Survey of Austria in 1984. To create a bioavailable Sr isoscape for the local region, a domain mapping approach was employed (Holt et al. 2021). Using the 87Sr/86Sr of all sampling sites within the geological unit, the median 87Sr/86Sr was calculated for each unit in the region using the Summarize Within tool in ArcGIS Pro 2.8.0. Using the median for each unit avoids influence from outlying measurements. Median Absolute Deviation (MAD) was calculated for each unit to highlight the errors associated with the isoscape prediction. To provide elevation, sampling sites were also spatially joined with a Digital Elevation Model (EU-DEM v1.1 2018).

A typo-chronological assessment of the finds combined with correspondence analysis established the chronological sequence of graves within the cemetery (Bølviken et al. 1982; Ringrose 1992) using Past 4.10 (Hammer et al. 2001) and CA-PCA analysis for Excel (Madsen 2016). To anchor the relative dating in an absolute chronological framework, ten calcined human bone elements were radiocarbon dated. The same bone elements as for strontium isotope analysis was selected to ensure the bones belong to the same individual within the grave (e.g. Sabaux et al. 2021). Radiocarbon data was modelled using OxCal 4.4 (Bronk Ramsey 2009) using the atmospheric data from Reimer et al. (2020). For further statistical analysis, the cemetery was divided into two phases based on the chronological assessment. To avoid an alpha-error accumulation due to multiple testing all p-values were corrected using the Bonferroni-Holm method (Holm 1979).

Results

Strontium isotope ratios of all human bone samples range from 0.7061 to 0.7124, with the highest variability seen in the diaphysis of the long bones (0.7061 to 0.7124). The petrous bone samples range from 0.7084 to 0.7113. The rib samples – a skeletal element with a higher turn-over rate compared to long bones – showed a narrower range than the other bone elements, with values ranging from 0.7085 to 0.7098. Using Tukey’s fence method based on a boxplot and Rosner’s Extreme Studentized Deviate test (generalized ESD test; p < 0.05) for identifying multiple outliers, five measurements were statistically determined to stand out within the cremated human bone samples (Rosner 1975; Tukey 1977). Among these, four individuals showed comparably higher and one a lower 87Sr/86Sr (Fig. 3). After excluding these five isotopic values, the widest range of 87Sr/86Sr was then observed in petrous bones (0.7084–0.7113), followed by diaphyses (0.7084–0.7102), and ribs (0.7085–0.7098).

Fig. 3
figure 3

87Sr/86Sr ratios of the long bone samples depicted in the box in relation to the BASr of the hills and the valley represented in a boxplot

Full size image

Several scenarios may explain the wide range of the plant samples used to construct the BASr map range between 0.7080 and 0.7162 The isotopic signature of the Lower Traisen Valley exhibits significant diversity, reflecting the geological complexity of the region. The measured values display a normal probability (Shapiro-Wilk test: p < 0.001, n = 163) and an even distribution. One notable observation is the stark contrast between the highest measurement (MFP134, 0.7162 87Sr/86Sr) and the subsequent measurements (MFP133, 0.7150 87Sr/86Sr, and MFP135, 0.7148 87Sr/86Sr) taken at the same location. The considerable difference of 0.0012 between these measurements suggests potential contamination or alteration in the sample or the site, making its interpretation respective the local value problematic. However, aside from this anomaly, the measured values support the assumption that they reflect the local strontium isotopic signature well. Thus, a local BASr range of 0.7080–0.7150 can be established.

The sampled region was further geographically divided into two distinct areas: a valley encompassing plains and gravel terraces, and a hill area. The division is delineated by elevation levels, with the valley area reaching a maximum elevation of 235.4 m and the hill area at a minimum of 259.1 m. Statistical analysis revealed a highly significant difference between the two areas (valley: n = 89, hill: n = 74; Mann-Whitney-U test: p < 0.001; Cohen’s d: -1.656, r = -0.638), facilitating the differentiation of 87Sr/86Sr ratios between valleys and hills. By utilizing one standard deviation from the mean of both areas (valley: 0.7087 ± 0.0006, hills: 0.7111 ± 0.0019), a delineation threshold of 0.7093 for distinguishing between valley and hill ratios can be established (Fig. 4).

Fig. 4
figure 4

87Sr/86Sr of the modern plant samples geographically divided in hill (orange) and valley (blue) samples

Full size image

The burial community, except for the 3 to 18-year-old juvenile with the lowest ratio (0.7061; Grave 29/individual II), falls within a range from 0.7083 to 0.7124 and covers most of the BASr spectrum; only the highest BASr values (above 0.7124) are not attained. The other four individuals statistically identified as outliers, still fall within the local BASr range.

Out of the fifteen petrous bone samples, fourteen showed little deviation from the long bone signals (between < 0.0001 and 0.0007 87Sr/86Sr difference). In the 30-50-year-old male individual from Grave 163, a pronounced discrepancy (0.0024 87Sr/86Sr difference) was observed between their strontium isotope values: both values by themselves fall within the established local range, but they differed substantially (petrous bone 0.7113, diaphysis 0.7090, Fig. 5). 87Sr/86Sr from 18 individuals, for which both long bone diaphysis and rib samples were measured, did not reveal relevant differences within each person (< 0.0001 to 0.0004 87Sr/86Sr difference), indicating no detectable migration signal during the last phase of their lives.

Differences between the values of different individuals found in multiple burials, except the non-local individual in Grave 29, did not exceed 0.0008 87Sr/86Sr difference. In 10 out of 19 cases differences less than 0.0001 in 87Sr/86Sr were observed.

After excluding graves and types of finds with too little chronological information the Correspondence Analysis still enveloped 136 burials. It was decided to include as many burials with 87Sr/86Sr as possible compromising on chronological significance (Porčić 2013; Theune 1995). The ten 14C samples spanned a range from 1446 BCE to 836 BCE and were used to test the Correspondence Analysis and differentiate two cemetery phases.

Fig. 5
figure 5

87Sr/86Sr of the petrous bone samples in relation to the associated long bone 87Sr/86Sr. Individual 163 shows a discrepancy of 0.0024 87Sr/86Sr indicating changing food sources during lifetime

Full size image

Discussion

The map

Since the Lower Traisen Valley is located within a geologically diverse area, where various major geological regions meet, the high density of sampling points ensures that all geological Sr variation in the region is captured. The wide range of strontium isotope values (87Sr/86Sr from 0.7080 to 0.7162) measured suggests this occurred. The Traisen Valley floor returned the lowest 87Sr/86Sr in the region, median of 0.7083 (n = 50). Gravels from the lower, higher, and older terraces along the valley edge returned higher 87Sr/86Sr median values (± MAD); 0.7089 ± 0.0001 (n = 21), 0.7096 ± 0.0001 (n = 6) and 0.7086 ± 0.0006 (n = 3) respectively. The hills surrounding the valley returned higher 87Sr/86Sr, with the Oncophora, Loess and Conglomerates layers showing 87Sr/86Sr from 0.7094 to 0.7122. To the west of the valley, the highest 87Sr/86Sr was found in the granulites layer (87Sr/86Sr = 0.7141 ± 0.0004, n = 9). Even within a relatively small radius of 10 km, notable variations in isotopic values, with differences as high as 0.0082, have been observed. These differences are particularly pronounced within specific geological units, especially along the valley slopes where geological outcrops and movements are prevalent (Fig. 6, B). Nevertheless, a well-defined local strontium spectrum (BASr 0.7080–0.7150 87Sr/86Sr) was established, as well as a division of the local area into two 87Sr/86Sr zones with a delineation threshold of 0.7093.

The detailed mapping of this region offers the possibility not only to answer questions about land and resource use but also to discuss them on an individual, social and community level. (Fig. 6). In Austria, the Traisen Valley is currently the only well-sampled landscape for BASr. Only a few other regions, such as the area around Stillfried, Lower Austria (Retzmann et al. 2020), and Kleinhadersdorf (Neugebauer-Maresch and Lenneis 2015), Lower Austria, have been sampled.

Fig. 6
figure 6

A Median 87Sr/86Sr for each sampling site and surface lithology, and B Median absolute deviation (MAD) for each surface lithology. The location of the Inzerdorf cemetery is indicated with a X and on B plant sampling sites indicated with black dots

Full size image

The population

The burial community, excluding the non-local individual, encompasses a 87Sr/86Sr range from 0.7083 to 0.7124 and spans a significant portion of the BASr range. Several scenarios may explain the wide range of 87Sr/86Sr seen in the individuals buried at Inzersdorf. First, if an individual is eating locally, its 87Sr/86Sr will gradually adapt to the local range due to the continuous remodelling of the bones, which may result in previously non-local individuals falling within the local range after a few years. However, the exact process and timeframe of remodelling specific bones and elements in relation to their 87Sr/86Sr are presently not well understood (Veselka et al. 2021a). Second, the individuals might be non-local but come from an area with similar geology resulting in similar isotope ratios. Third, the individuals may have consumed food grown away from the main food production area of the burial community. The hilly regions with relatively high 87Sr/86Sr were most likely used for food procurement. These individuals may have lived their life just a little removed, maybe in single farmsteads in the hills, rather than within the main settlement, if such a thing can be assumed. This would further imply that the burial ground of Inzersdorf was not exclusively used by a single geographically and socially enclosed community, but the communal burial ground for several farms. However, values exceeding 0.7124, which represent the highest BASr levels in the region, are not observed. This is not surprising, since such high values are found in the hills on the opposite side of the river Traisen and the periphery of the Bohemian Massif of the Dunkelsteiner Forest (Fig. 6). It can be inferred that large-scale food procurement did not occur in these areas, whereas the hills west of the Traisen and the riverbanks likely represent the locations where the population obtained their food from. Late Bronze Age settlement traces were documented frequently on both sides of the Traisen Valley, including near the cemetery (Adametz 2011; Adametz and Lochner 2021). Due to the lack of investigation and publication on the specific sites, a direct relationship between one or more settlements with the cemetery of Inzersdorf cannot be concluded, although the human strontium data suggest a left bank location.

Skeletal elements – age - sex

Excluding statistical outliers from the main population, the broadest range of 87Sr/86Sr was observed in petrous bones (0.7084–0.7113, SD: 0.0008, n = 15), followed by diaphyses (0.7084–0.7102 SD: 0.0006, n = 177) and ribs (0.7085–0.7098, SD: 0.0004, n = 18). If only the 87Sr/86Sr of individuals with more than one measurement are taken into account, a similar picture emerges (petrous bones: 0.7084–0.7113, SD: 0.0008, n = 15, diaphysis: 0.7084–0.7101, SD: 0.0005, n = 18, rib: 0.7085–0.7098, SD: 0.0004, n = 18). This decrease in variation between different skeletal elements demonstrates a gradual adaptation to a more specific local value during individuals’ lifetimes. The different strontium isotope ranges of the varying skeletal elements suggest a greater variation in the geographical origin of the individuals or food sources during earlier life and a more settled lifestyle during the later years. The wider variety of values seen in the petrous bones, which represents both the mother’s diet during pregnancy and breastfeeding, and the early childhood diet, and the smaller range of 87Sr/86Sr observed in the older age group further support the idea of an age-related food procurement. Likewise, females exhibit a wider range of 87Sr/86Sr compared to males, indicating similar, but broader food sources for some females and more restricted food procurement strategies for males.

Multiple burials

87Sr/86Sr measured in burials with two or more individuals generally span the entire local BASr range, except for the double burial in Grave 29, which contained the non-local juvenile with a very low value of 0.7061. However, the differences between the values were relatively low (≤ 0.0008, 10 out of 19 cases less than 0.0001, Fig. 7) These findings suggest that co-buried individuals generally ate food growing on a similar geological substrate indicating similarity in lifestyle between them.

Fig. 7
figure 7

The general small 87Sr/86Sr variation within co-buried individuals within multiple burials (represented as one bar) indicates a similarity in lifestyle of co-buried individuals

Full size image

Chronological aspects

To examine the relationship between 87Sr/86Sr and cemetery chronology, a correspondence analysis of the graves was established that aimed to include as many burials with 87Sr/86Sr results as possible (Supplementary Material).This approach retains graves with minimal equipment, such as many child graves, and maximises the coverage across all social groups, but compromises on chronological significance (Porčić 2013; Theune 1995). As a result, the cemetery was divided into two phases based on the seriation of artefact and grave types; the ten radiocarbon dates match the established sequence (Figs. 8 and 9).

Fig. 8
figure 8

Calibrated radiocarbon dates used to test the correspondence analyses. The different colours indicate the two phases

Full size image
Fig. 9
figure 9

Correspondence analysis of the graves of Inzersdorf. The radiocarbon dated graves are highlighted and the varying colours indicate the two phases

Full size image

Dividing the strontium data measured from the diaphysis into an earlier (approx. 1372 BCE − 1100 BCE) and a later phase (approx. 1100 BCE − 920 BCE) based on the seriation revealed a significant difference, with higher mean 87Sr/86Sr found in the earlier compared to the later phase (earlier phase: n = 44, later phase: n = 45; Mann-Whitney-U test: p < 0.001; Cohen’s d: 10,670, r = 0.983). This difference is even more pronounced as a gap appears when only data based on 14C results were considered (no overlap, 0.0004 difference in 87Sr/86Sr between the earlier and the later phase, (earlier phase: n = 5, later phase: n = 5; Mann-Whitney-U test: p = 0.009), minimizing chronological uncertainties Fig. 10).

Comparing 87Sr/86Sr of the earlier and later phases of the cemetery with the baseline reveals a wide distribution of 87Sr/86Sr encompassing both the valley plains and the hills, with more values falling into the hill range. During the later phase, the range of 87Sr/86Sr becomes considerably narrower (0.7083 to 0.7099), and comparable to the valley plains and lowermost gravel terraces. This suggests that during the earlier phase, food was primarily procured in the hilly areas, and to a lesser extent in the valley plain, whereas during the later phase, agricultural produce was mainly obtained from the valley plains. An assessment regarding whether the broader range of values observed in the earlier phase of the cemetery could be attributed to the founder effect, where founding individuals come from diverse regions, has been deferred. This decision is primarily due to the extended duration of the older phase of the cemetery. Further subdivision into smaller phases would result in an insufficient number of individuals with available strontium values, thereby impeding meaningful statistical analysis.

Fig. 10
figure 10

87Sr/86Sr of the diaphysis samples divided into the earlier and later phase based on the correspondence analysis (left) and 14C dates (right) depicted in a boxplot

Full size image

The chronological difference in 87Sr/86Sr is apparent in both female and male individuals (males: earlier phase: n = 8, later phase: n = 14; Mann-Whitney-U test p < 0.001; females: earlier phase: n = 16, later phase: n = 17; Mann-Whitney-U test: p < 0.001) with greater differences in the male group (Cohen’s D females: 1.1879, r = 0.685; males: 2.945, r = 0.827). In juveniles, chronological differences were not as pronounced as a large overlap can be seen while the effect strength is high (earlier phase: n = 22, later phase: n = 18; Mann-Whitney-U test p = 0.004; Cohen’s D = 4.552, r = 0.916), perhaps because they are less often equipped with significant grave goods and more difficult to date (Fig. 11).

Fig. 11
figure 11

Gender differences in 87Sr/86Sr in relation to the phases of the cemetery depicted in a boxplot

Full size image

In summary, it can be inferred that all social groups followed the general trend of the burial community during both phases of the cemetery, but women and children had access to a wider range of food sources compared to men, whose nutrition sources appear to have been more limited. While increased mobility of women and juvenile individuals may explain this trend, alternative food procurement strategies must be taken into consideration.

It has been argued that long rectangular graves are the oldest in the cemetery because they show close similarities with the preceding Middle Bronze Age (Lochner 2013). Further, it was inferred that the rapid change to the practice of urn burials might have been introduced by newcomers (Cavazutti et al. 2022). To test this theory, the strontium values of the long rectangular graves were compared with those from the round graves and a statistical significance was identified (long rectangular graves: n = 12, round graves: n = 171; Mann-Whitney-U test p = 0.001). However, this difference is related to the chronological change in land use and cannot be maintained if only the round graves of the earlier phase are compared to the long rectangular ones (long rectangular graves: n = 12, round graves EPH: n = 48; Mann-Whitney-U test p = 0.18).

Locals and non-locals: a travelling metalworker at Inzersdorf?

Five 87Sr/86Sr values were considered outlying data points in statistical tests. The subadult (age based on epiphyseal closure and long bone diameter; individual 029/II) from the double burial within Grave 29 with 87Sr/86Sr of 0.7061 falls below the local range and can be identified as a non-local individual. The other individual in the grave, an adult male aged 20–40 years at death, fits well with the local range (87Sr/86Sr = 0.7098). Double burials are rare, but not unknown in Late Bronze Age urnfields, but the fact that both individuals were buried within separate urns makes this grave unique at Inzersdorf. Due to the poor preservation of the grave, not all artefacts are precisely locatable within the grave, including some fire-affected lithic artefacts and various ceramic and bronze fragments. The 3-18-year-old was equipped with a casting mould, the only one at the cemetery and unusual as a grave good (Fritzl 2017), indicating a connection to bronze casting and metal working. Similar moulds have been found in high numbers nearby (Neugebauer 1994), primarily from settlement contexts (Overbeck 2018). Other than that, the urn contained an unfinished bronze workpiece, a spindle head pin and further lithics. The co-buried individual was equipped with an awl, arm ring fragments, further bronze fragments and a small cup (Fig. 12). Due to the relatively nondescriptive archaeological material, typology cannot be used to precisely date the grave. The 14C date hints at the earlier phase to the beginning of the later phase of the cemetery (1201–1008 calBCE, 95.4% probability). Further, none of the artefacts hint at a specific area of origin.

Fig. 12
figure 12

Equipment of grave 029

Full size image

Several copper mines were used during the Bronze Age in Europe, for example in Trentino, Italy, or Špania Dolina at the northern rim of the Carpathian Basin (Radivojević et al. 2019). However, the closest copper sources exploited at this time are found in the eastern Alps, especially in the Mitterberg region, the Eisenerz Alps, and the Semmering region (Klemm and Trebsche 2021; Pernicka et al. 2016).

The exceptionally low 87Sr/86Sr of individual 029/II, however, suggests a geologically young bedrock from which the young person’s diet had been sourced. Over the last decade, predictions of 87Sr/86Sr based on the geological subsurface have been made using different sampling and mapping techniques; while promising, it is generally agreed that small-scale geological and BASr variations remain challenging (Bataille et al. 2018; Hoogewerff et al. 2019; James et al. 2022; Lugli et al. 2022). Comparably low 87Sr/86Sr are found only in regions associated with volcanic activity such as the basalt formations in north-eastern Ireland (Snoeck et al. 2016, 2020), around the Italian Etna (Lugli et al. 2022) or in France (Willmes et al. 2018). The closest region with past volcanic activity is southern Styria and southern Burgenland (Kellerer-Pirklbauer and Fritz 2022).

The easternmost Alpine copper mines in western Styria (Klemm and Trebsche 2021) combine copper sources and bedrock generated by volcanic activity. We speculate that individual 029/II originated from this region, based on the casting mould as a grave good and the remarkably low Sr isotope ratio. However, other possible origins should not be dismissed, since the wider geographical context includes regions such as the western Hungarian plains, the Banská Bystrica region in western Slovakia, north-eastern Bohemia, or Trentino-Alto Adige (https://geoportal.bgr.de/mapapps/resources/apps/geoportal) with similarly young geology and possible copper sources.

Smiths and other metalworkers are commonly thought to have been of high social status (Bartelheim 2009; Harding 2000), but aside from the mould, no grave equipment suggests a high social position for individual 029/II. However, the combination of a non-local isotope ratio and bronze working equipment might hint at bronze-related mobility, connected to raw material procurement, craftworking or knowledge transfer as the young age at death could indicate a student or apprentice to a bronze trader or travelling smith.

The four humans with higher 87Sr/86Sr fall within the local BASr range. The only long-rectangular Grave 88 contained the scattered cremated bones (2.15 g) of a three to nine-year-old child, with the highest 87Sr/86Sr in the cemetery (0.7124). The child was buried with the remains of five ceramic vessels, a bronze pin and fragments of spiral rolls. Grave 88 belongs to the earlier phase of the cemetery. Grave 116 held 12.71 g of cremated remains of a 9.5 to14.4-year-old with an 87Sr/86Sr of 0.7115, Grave 153 included 96.12 g of cremated remains of a 7 to12-year-old with a very similar 87Sr/86Sr (0.7114). Neither of the scattered cremation burials had grave goods, which is common for children’s graves but makes a chronological assessment difficult. The 14-to-19-year-old individual from Grave 260 (0.7109) was represented by 151.66 g of cremated remains and several fire-affected vessels that hint at the early phase of the cemetery which is supported by the radiocarbon date (1271 − 1056 calBCE, 95.4% probability) and represents average equipment for juvenile individuals (Fig. 13).

Fig. 13
figure 13

Equipment of grave 088, 116, 153 and 260

Full size image

The male individual in Grave 163, who died between the ages of 30 and 50 years, is noteworthy because the 87Sr/86Sr of the petrous bone (0.7113) and the diaphysis (0.7090) differed (0.0023). The petrous bone 87Sr/86Sr, the childhood signal, falls within the range of the hills, while the diaphysis ratio more closely aligns with the valley. This suggests either movement or a change in food sources, from production areas in the hills to those nearer to the valley plain. The change of food source most likely occurred between childhood and the last third of the individual’s life (between 2 and 30 years). After death, the individual was buried in an urn, and deposited with three funerary vessels, the left foreleg of a sheep/goat, and fragments of three other vessels. In the context of the cemetery, this constitutes an average to well-equipped grave (Fig. 14). The grave belongs to the late phase of the cemetery.

Fig. 14
figure 14

Equipment of grave 163

Full size image

The buried person appears to have been well integrated, as no differences to the main community in the grave goods were observed. Between the age of changing food sources and the age of death, a long time passed, suggesting the man had lived much longer with the main group and consequently had more opportunities to socially adjust. Another possible scenario would be, that the population stopped using the hilly areas for food production during his lifetime. The chronological position of the grave fits this assumption as the change in land use roughly coincides.

All statistically outlying 87Sr/86Sr measurements represent childhood data, ranging approximately from 7 to 19 years at the time of death; or, as individual 163 that shows a changing food source during childhood or early adulthood. This suggests a higher level of mobility and/or a different food source for a subset (5 of 105 tested (118 subadults in total), 5%) of the juvenile population compared to the other age groups. Similar results have been observed in other studies (Grupe et al. 2017; Knipper et al. 2018). The values are consistent with those observed in the hill area surrounding the Traisen, suggesting that the individuals may have relocated from surrounding farmsteads or other nearby settlements to the main village to which the burial ground belonged during their later childhood. Alternative explanations include dietary changes to different sources of nutrition during development, intense involvement in activities requiring short-term mobility such as herding, which has traditionally been regarded as child labour in many societies (Food and Agriculture Organization of the United Nations 2013), or the return transportation of cremated children’s remains for burial in the ancestral burial ground (Ensor 2021).

Conclusion

As the examined data shows, the only factual non-local individual must be considered separately. The related metalworking equipment hints at bronze-related travel, such as raw material procurement and distribution, trade with objects or knowledge distribution. The combination of isotopic value, grave goods and construction suggest a respected position, though probably not one of the highest, for this metal specialist.

Further, it can be concluded that if evidence for mobility can be seen in the deviating strontium data, subadults of all social strata would be the main mobile community members, probably contributing to the socio-cultural and socio-economic dynamics.

The high resolution and variability within both the plant and human isotope data allowed differentiation between primarily valley-based versus hill-based diets. All social groups followed the general trend, although women and children had access to food originating from more varied locations compared to men.

Relating this result to the chronology of the cemetery, a diachronic shift in land use can be observed. While in the earlier phase of the cemetery food production on the hills and also to a lesser extent in the valley plains occurred. Later food production almost exclusively took place in the valley plains. Further, the earlier phase of the cemetery shows a higher diversity in food procurement especially for women and children, which may indicate greater small-scale regional mobility for them but might also indicate the common usage of the burial grounds for people living in different areas. The narrow concentration with almost no exceptions on the plains in the later period suggests a more specialised, enclosed and deeply established community, especially for the male population. Therefore, connectivity seems to predominantly have happened through networking, trade and social ties rather than migration and long-term mobility.

These results highlight the importance of exploring indirect evidence for land use, such as through strontium isotope analysis, particularly since direct evidence such as gardens, fields and herding areas are hard to identify, even if the terrain is not as highly disturbed by human activity as it is in the Traisen Valley.

Further, results demonstrate that extremely poorly represented individuals (less than 20 g of cremated remains, such as Grave 88 and 116), which appear limiting from a traditional anthropological view, can provide much information through their strontium isotope values. Finally, Grave 163 highlights the advantage of comparing several measurements of isotope values within one individual, as the observed changing nutrition indicates a shift in lifestyle during childhood or early adulthood.

There was relatively little evidence of mobility within the burial community of the cemetery overall, taking into account the diverse geology of the region and the mixed signals from remodeled bones (end-of-life signals). Thus, it seems that the cultural development of the Late Bronze Age in this area was not primarily influenced by migration, but probably by other factors.