Introduction
Throughout Northwestern Europe, from the Neolithic to the Early Middle Ages, two main funerary rituals took place: cremation and inhumation (Capuzzo et al. Reference Capuzzo, Snoeck, Boudin, Dalle, Annaert, Hlad, Kontopoulos, Sabaux, Salesse, Sengeløv, Stamataki, Veselka, Warmenbol, De Mulder, Tys and Vercauteren2020, Reference Capuzzo, De Mulder, Sabaux, Dalle, Boudin, Annaert, Hlad, Salesse, Sengeløv, Stamataki, Veselka, Warmenbol, Snoeck and Vercauteren2023; Rebay-Salisbury Reference Rebay-Salisbury2017). They sometimes co-occurred or happened successively within the same region, or even on the same burial ground (e.g. Lippok Reference Lippok2020; Snoeck et al. Reference Snoeck, Pouncett, Ramsey, Meighan, Mattielli, Goderis, Lee-Thorp and Schulting2016, Reference Snoeck, Jones, Pouncett, Goderis, Claeys Ph, Zazzo, Reimer, Lee-Thorp and Schulting2020; van den Broeke Reference van den Broeke2014; Veselka et al. Reference Veselka, Capuzzo, Annaert, Mattielli, Boudin, Dalle, Hlad, Sabaux, Salesse, Sengeløv, Stamataki, Tys, Vercauteren, Warmenbol, De Mulder and Snoeck2021a). The reasons behind the choice of one ritual instead of the other are still poorly understood and widely debated (e.g. Lippok Reference Lippok2020; Rebay-Salisbury Reference Rebay-Salisbury2017). The study of cremated remains is rendered even more complex due to the highly fragmented nature and often incompleteness of the human remains found in such burials. Indeed, cremation was the main funerary practice in the Belgian Meuse valley during the Late Bronze (LBA)–Early Iron Age (EIA) and due to the fragmentary aspect of the human remains, cremation graves were either not studied or the older studies provide limited information, highlighting the need for new and in-depth research (De Mulder Reference Louwen2011; Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021; Stamataki et al. Reference Stamataki, Kontopoulos, Salesse, McMillan, Veselka, Sabaux, Annaert, Boudin, Capuzzo, Claeys, Dalle, Hlad, Sengeløv, Vercauteren, Warmenbol, Tys, De Mulder and Snoeck2021).
With just osteological analysis of the bone material within a cremation deposit, it is not possible to ensure that all the human bone fragments in the grave belong to a single individual. Even if ageing and sexing methods have been adjusted and adapted to cremated remains in the past years (e.g. Hlad et al. Reference Hlad, Veselka, Steadman, Herregods, Elskens, Annaert, Boudin, Capuzzo, Dalle, De Mulder, Sabaux, Salesse, Sengeløv, Stamataki, Vercauteren, Warmenbol, Tys and Snoeck2021; Veselka et al. Reference Veselka, Hlad, Steadman, Annaert, Boudin, Capuzzo, Dalle, Kontopoulos, De Mulder, Sabaux, Salesse, Sengeløv, Stamataki, Vercauteren, Tys and Snoeck2021b), the Minimum Number of Individuals (MNI) is still commonly based on the presence of multiple unique skeletal elements and differences in bone robusticity (e.g. adult vs. nonadult) (Albanese et al. Reference Albanese, Cardoso and Saunders2005; Cavazzuti et al. Reference Cavazzuti, Bresadola, D’Innocenzo, Interlando and Sperduti2019; Gejvall Reference Gejvall, Brothwell and Higgs1963; Gonçalves Reference Gonçalves2011; Gonçalves et al. Reference Gonçalves, Thompson and Cunha2013; Osipov et al. Reference Osipov, Harvati, Nathena, Spanakis, Karantanas and Kranioti2013; Thompson Reference Thompson2002; Van Vark et al. Reference Van Vark, Amesz-Voorhoeve and Cuijpers1996). However, those elements are often absent making the MNI of cremation deposits imprecise.
In the early 2000s, it has been shown that reliable radiocarbon dates could be obtained from fully calcined bones (Lanting et al. Reference Lanting, Aerts-Bijma and van der Plicht2001; Van Strydonck et al. Reference Van Strydonck, Boudin and De Mulder2010; Reference Van Strydonck, Boudin, Hoefkens and de Mulder2005). In addition, it was also demonstrated that calcined bone provides a reliable substrate for strontium isotopes (87Sr/86Sr) analysis in archaeological contexts (Snoeck et al. Reference Snoeck, Lee-Thorp, Schulting, de Jong, Debouge and Mattielli2015, Reference Snoeck, Pouncett, Claeys Ph, Mattielli, Parker Pearson, Willis, Zazzo, Lee-Thorp and Schulting2018) and strontium isotope analyses are now widely applied across Europe (Snoeck et al. Reference Snoeck, Cheung, Griffith, James and Salesse2022). A recent study in Belgium on the LBA site of Herstal “Pré Wigier” (Figure 1) combining 87Sr/86Sr with radiocarbon dating highlighted higher MNI in two of the graves than defined by the osteological study alone (Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021). In grave 4, two individuals were identified by combining 87Sr/86Sr and radiocarbon dating while only one was determined by osteological analysis. Similarly, in grave 6, two individuals were determined osteologically, but the combination of radiocarbon dates and 87Sr/86Sr showed that at least three individuals were present. Some of the calcined diaphyses found in these two graves had indeed quite different 87Sr/86Sr (> 0.0010) and thus unlikely belong to the same individual. Furthermore, they returned radiocarbon dates up to more than a century apart. These three individuals from Grave 6 of Herstal were all nonadults, while in general, nonadults are associated with adults in LBA–EIA plural burials (Larentis Reference Larentis2023; Sørensen and Rebay-Salisbury Reference Sørensen and Rebay-Salisbury2023). The presence of different individuals belonging to different time periods within a single cremation deposit raises the issue of intentionality and potentially of bone curation (Booth and Brück Reference Booth and Brück2020; Brück and Booth Reference Brück and Booth2022; Rebay-Salisbury Reference Rebay-Salisbury2018).
The location of the studied LBA–EIA sites in the Meuse Valley, Belgium.
The aim of this paper is twofold: first, understand if the phenomenon of multiple deposits containing bones belonging to different chronological events in a single cremation grave appears elsewhere in the Meuse valley, Belgium, or if it is unique to Herstal. Second, demonstrate how combining osteological analyses with radiocarbon dates and 87Sr/86Sr improves the assessment of the MNI in cremated graves and how the method can be used on a larger scale. This study also highlights the importance of carrying out 87Sr/86Sr analyses and radiocarbon dating on the same bone fragments to enable direct comparison of the 87Sr/86Sr results and the radiocarbon dates.
Material and methods
For this research, graves 4 and 6 of Herstal “Pré Wigier” studied in Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021) were reanalyzed, additional samples were taken from grave 4, and four graves with separate bone deposits from three additional sites from the Meuse Basin were included (Figure 1; Table 1). Herstal, near Liège, was excavated between 1965 and 1966 and presents 21 graves from the LBA to the EIA transition (Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021). Grave 6 consisted of a small pack of bones on which an urn containing bones and charcoal was deposited (Alenus-Lecerf Reference Alenus-Lecerf1974). Grave 4 contained an urn filled with cremated remains and in which an accessory vase also containing several bone fragments was buried (Alenus-Lecerf Reference Alenus-Lecerf1974). The site of “Grand Bois”, near the village of Saint-Vincent, located in the province of Luxembourg, was excavated from 1882 to 1964 during several excavation campaigns. In this LBA–EIA site, at least 88 tumuli with cremation graves were found. Two of those graves, 23 and 41, present double bone deposits. Grave 23 contained two urns with cremated bones and few other bone fragments were present on the surface of the pyre (Mariën Reference Mariën1964). According to the author, the urns were placed in the grave simultaneously. Grave 41 contained an urn filled with cremated bones and charcoal in and around it. The urn was deposited on the pyre and was covered by a tumulus of 4 m diameter. Several cremated bone fragments were still scattered on this pyre (Mariën Reference Mariën1964). Grave 9 from the LBA site of “Achelse Dijk” in Neerpelt, in the Limburg province, was excavated in 1962–64. The cremation deposit seems to be a block of bones divided into two deposits (9t4 and 9u4) (Roosens et al. Reference Roosens, Beex and Van Impe1975). The last analyzed grave, 85-143, belongs to the site of Rekem “Hangveld”, also in the Limburg province. This BA–Iron Age (IA) cemetery contains more than 236 cremation deposits and was excavated in the 20th century. Grave 85-143 presents a biconical urn filled with cremated bones and few fragments of charcoal (Temmerman Reference Temmerman2007; Van Impe Reference Van Impe1980).
Typology and description of each deposit per grave
The methods for typology, osteoarchaeological analysis, 87Sr/86Sr analyses and radiocarbon dating used in the study are fully described in Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021) and can be found in SI. The six cremation graves were characterized using the archaeological reports of the sites and the typological classification of De Mulder (Reference Louwen2011). De Mulder’s funerary urn typology (Reference Louwen2011) has been created to describe the cremation burials in the Scheldt valley and was successfully applied to the study of Herstal (Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021). Since those six graves had two bone deposits each (A and B), both deposits were sampled. In total, between two and sixteen bone fragments were taken per grave for 87Sr/86Sr analyses. All 87Sr/86Sr were carried out at the Université Libre de Bruxelles (ULB) and the Vrije Universiteit Brussel (VUB) following Snoeck et al. (Reference Snoeck, Lee-Thorp, Schulting, de Jong, Debouge and Mattielli2015). In total, 62 calcined bone fragments from the six graves were used (19 samples from Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021) plus 43 additional samples. From each cremation deposit, if available, diaphyseal, rib, and cranial fragments were selected to take bone turnover into account as ribs have a faster remodelling rate than diaphysis (Clarke Reference Clarke2008; Fahy et al. Reference Fahy, Deter, Pitfield, Miszkiewicz and Mahoney2017). A difference in strontium isotope ratios between two diaphysis or two rib samples (Δ87Sr/86Sr) larger than 0.0010 is taken here as a threshold to consider two diaphysis or two rib fragments to belong to two distinct individuals. Nevertheless, this value has to be taken with cautions as Sr turnover rates in bone are still poorly understood and the variations between bones of the same individual also depends on the variability in 87Sr/86Sr of the different foods available to a particular individual.
Radiocarbon dates were performed on 19 calcined bone fragments selected amongst those used for 87Sr/86Sr analyses (i.e. 87Sr/86Sr and radiocarbon dates were obtained from the same bone fragments). The radiocarbon dating procedure followed the Royal Institute for Cultural Heritage (KIK-IRPA, Brussels, Belgium) protocol (Boudin et al. Reference Boudin, Van Strydonck, Van Den Brande, Synal and Wacker2015; Van Strydonck et al. Reference Van Strydonck, Boudin, Hoefkens and de Mulder2005, Reference Van Strydonck, Boudin and De Mulder2009; Wojcieszak et al. Reference Wojcieszak, van den Brande, Ligovich and Boudin2020). In total, the radiocarbon dates of 26 calcined bone fragments (7 dates from Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021 and 19 additional dates) from six different graves (Herstal 4 and 6, Achelse Dijk 9, Grand Bois 23 and 41, Rekem 85-143) were calibrated using the software OxCal 4.4 (Bronk Ramsey Reference Bronk Ramsey2009) and the IntCal20 calibration curve (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Ramsey, Butzin, Cheng, Edwards and Friedrich2020). Multiple radiocarbon measurements from the same graves were statistically tested for contemporaneity using the R_Combine tool in OxCal 4.4, providing information on the consistency of the time series via a χ2-test (Ward and Wilson Reference Ward and Wilson1978).
Results
To categorize the different graves, the main deposit (deposit A) of each grave was compared to De Mulder’s typology (Table 1). Graves 4 and 6 of Herstal were characterized respectively as type A: urn graves containing burnt bone material and possible grave goods and type B: urn containing burnt bones and pyre remains, such as charcoal (Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021). Achelse Dijk is defined based on the archaeological survey from the report as a type C, as “bone pack grave”, bundled burnt bone material suggestive of the use of an organic container, while graves 23 and 41 of Grand Bois are type H, bone selected and deposited in an urn on the pyre location, according to their description. The grave 85-143 is defined as a type A (Temmerman Reference Temmerman2007).
The total weight of the cremations varies between 28.3 and 1172.0 grams, and 4 out of 6 graves contained animal bones (Table 2). The osteological results of Herstal identified a MNI of one individual in grave 4 and two individuals in grave 6 based on the presence of two right petrous part and differences in dental age (Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021). For Grand Bois 23, one adult of more than 18 years based on the vertebral rims fusion is identified and for 41, one nonadult between 13 and 18 years old based on the unfused coracoid process of the scapula. The individual of grave 9 from Achelse Dijk could not be identified due to the small amount and high fragmentary state of the bone material. Rekem 85-143 might present two individuals: one adult between 19 and 40 years and another of unknown age according to Temmerman (Reference Temmerman2007). If the weights per deposit of each grave are compared, the highest weight is in 67% of the cases (4 out of 6) associated with deposit related to the principal container.
Osteological results
* Data from Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021); M = male, NA = not applicable, I = indeterminate.
The calibrated radiocarbon dates (Table 3, Figure 2) from the cremation deposits range from LBA to EIA, from the 12th to the 6th century BC. Large calibrated confidence intervals have been obtained for the dates falling on the Hallstatt plateau, between 800 and 400 BC in the IntCal20 calibration curve which was caused by variation in solar activity (Pearson et al. Reference Pearson, Pilcher and Baillie1983; Stuiver and Becker Reference Stuiver and Becker1986; Wijma et al. Reference Wijma, Aerts, van der Plicht and Zondervan1996; van der Plicht Reference van der Plicht, Scott, Alekseev and Zaitseva2004). Performing a χ2-test permits to obtain more precise and accurate radiocarbon dates with narrower confidence intervals and to detect if two samples from the same grave belong to the same chronological event or not (Ward and Wilson Reference Ward and Wilson1978). It is important, however, to keep in mind that because of a potential old-wood effect two bone fragments that do not belong to the same chronological event, based on the radiocarbon dates, could still belong to the same individual (see discussion).
Radiocarbon dates and 87Sr/86Sr of the calcined human bones
* Data from Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021);
⸆ New radiocarbon date but Sr data from Sabaux et al. (Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021).
Calibrated radiocarbon dates with 87Sr/86Sr isotope ratios from Herstal, Achelse Dijk, Grand Bois and Rekem. Each colour in each grave corresponds to a different individual based on the combination of osteology, radiocarbon dating and strontium (MNI = 14 or 15) (*data from Sabaux et al. Reference Sabaux, Veselka, Capuzzo, Snoeck, Sengeløv, Hlad, Warmenbol, Stamataki, Boudin, Annaert, Dalle, Salesse, Debaille, Tys, Vercauteren and De Mulder2021).
For grave 4 of Herstal, the χ2-test of the dates with ID 08055 and 08066 in deposit A were too far apart to belong to the same chronological event (T = 7.5; 5% = 3.8; df = 1), with 08055 being between 64 and 185 years older than 08066 for 1σ probability, calculated using the tool Difference in OxCal 4.4. The same phenomenon was observed in grave 6 (T = 20.3; 5% = 9.5; df = 4), permitting to identify the radiocarbon dates belonging to at least two distinct chronological events. Three new dates were performed on samples from grave 4 and confirmed the presence of radiocarbon dates belonging to distinct chronological events (Table S1), samples 08458 and 08100 returning dates close to sample 08066 (T = 0.2; 5% = 6.0; df = 2) while 08108 returning dates closer to 08055 (T = 0.7; 5% = 3.8, df = 1). Those new samples are indeed too far apart to belong to the same event (08458, 08100, 08108 – T = 21.891; 5% = 6.0; df = 2).
Radiocarbon dates from grave 9 from Achelse Dijk cover a large time span between 257 and 671 years for the 1σ probability, calculated using the tool Difference in OxCal 4.4. (179–1201 years for the 2σ probability). The χ2-test shows that the dates (08460, 01329, 08461, 01328, 08462) do not correspond to the same chronological event (T = 185.995; 5% = 9.5; df = 4). The tests further revealed that at least three distinct chronological events are represented. 08460 and 01329 returning similar 14C dates (T = 1.8; 5% = 3.8; df = 1), 08461 being different to all other dates (see table 3), and 01328 and 08462 returning similar dates as well (T = 1.0; 5% = 3.8; df = 1). The grave was a pack of bones split in two deposit (A and B—Table 1) and the three bones with the oldest radiocarbon dates came from deposit A, while the fragments with the most recent dates came from deposit B (see Figure 2).
Graves 23 and 41 from Grand Bois present in both case one sample in deposit A that is older than the other ones. Indeed sample 04226 (T = 23.634; 5% = 9.5; df = 4) in grave 23 stands alone while samples 08464, 08463, 04227, 05097 seem to belong to the same time event (T = 3.5; 5% = 7.8; df = 3). The same pattern is identified in grave 41, 04224 is isolated (T = 52.044; 5% = 7.8; df = 3) while 08192, 08191, 04225 return similar dates (T = 0.8; 5% = 6; df = 2). Eventually, for grave 85-143 from Rekem, the χ2-test indicated that the dates with ID 08381 (deposit A) and 08382 (deposit B) are too far apart and do not belong to the same chronological event (T = 7.978; 5% = 3.8; df = 1).
The 87Sr/ 86Sr from the different bone fragments of the six graves range from 0.7088 to 0.7147 (Table 3, Figure 3). The 87Sr/86Sr differences between diaphysis of the same graves are quite large and range from 0.0003 up to 0.0013 for Herstal Grave 4, equals 0.0014 in Herstal grave 6, range from 0.0009 to 0.0012 for Grand Bois grave 23, from 0.0010 to 0.0021 for Grand Bois grave 41 and from 0.0012 to 0.0059 for Achelse Dijk. In Rekem grave 85-143, however, the results are very similar (0.7123 and 0.7125).
87Sr/86Sr from the graves with deposits A and B that showed large 87Sr/86Sr differences between the various skeletal elements from Herstal, Achelse Dijk, Grand Bois and Rekem individuals. Orange colour in each grave corresponds to the dated samples.
Discussion
Based on the different analyses carried out on the six cremation deposits different MNI can be calculated (Table 4). The osteological evaluation allowed to identify up to two individuals in Herstal 6 and Rekem 85-143 but could not detect the presence of additional individuals in the other graves. This is not altogether surprising seeing the extreme complexity of working with highly fragmented cremated bones. The strontium data, on the other hand, highlights differences between different bone fragments as high as 0.0059. Those differences could be the results of mobility between two regions with very different biologically available strontium. Indeed, the difference in 87Sr/86Sr might be explained by bone turnover as different bones represent different stages of life. However, the analysis of different bone fragments from the same skeletal category (i.e. with comparable turnover rates), still shows a large difference. An alternative explanation would then be, when two rib or two diaphysis fragments present large differences (Δ87Sr/86Sr > 0.0010), that these bone fragments belong to distinct individuals who consumed foods from different geological areas. The threshold of > 0.0010 is rather arbitrary and should be considered with caution and in relation with the biologically available strontium variability around the site under investigation. Still, in the case of Achelse Dijk 9, where a difference of 0.0059 is observed between two fragments of diaphysis, one can be quite confident that these bone fragments do not belong to the same individual.
MNI evaluated based on individual and combined methodologies
While 87Sr/86Sr provide additional data to detect distinct individuals within a cremation deposit, radiocarbon dates are a powerful tool to further investigate such graves. Two main possibilities can explain large differences in radiocarbon dates of burned bones: (1) the bones belong to different individuals or (2) the bones were affected by the old-wood effect. The measurements of multiple bones per graves enable, in most cases, to differentiate between these two options. Indeed, old-wood effect is not expected to be homogenous as experimental studies have shown that cremated bone can incorporate between 36 and 95% of carbon from the fuel (Hüls et al. Reference Hüls, Erlenkeuser, Nadeau, Grootes and Andersen2010; Rose et al. Reference Rose, Meadows and Henriksen2020; Snoeck et al. Reference Snoeck, Brock and Schulting2014; Zazzo et al. Reference Zazzo, Saliège, Lebon, Lepetz and Moreau2012). As such, if there is a large variability between the dates, old-wood effect is a possibility. This is observed in the five dates of Achelse Dijk 9. In Rekem 85-143, as only two bone fragments were radiocarbon dated, it is also difficult to exclude the old-wood effect. In all other cases, however, two groups of dates are observed, suggesting two different cremation events, and thus two different individuals.
It is when combining all three methodologies (i.e. osteology, radiocarbon dating and strontium isotope analyses) that this method becomes really powerful in identifying the MNI of a cremation grave (Table 4). In the case of Achelse Dijk 9, the radiocarbon dates suggest a potential old-wood effect as an explanation for the large spread in the dates. However, the strontium isotope data clearly shows that the intermediate date (grey in Figure 2) must belong to another individual seeing the very large difference in 87Sr/86Sr between the five diaphysis fragments (0.0059). As such, there is not one but three individuals in that grave. In Rekem 85-143, the two bone fragments have very similar 87Sr/86Sr (0.7123 and 0.7125) and are not different enough to confirm the presence of two individuals.
In Grand Bois 23, 14 bones were analyzed for 87Sr/86Sr of which 5 were radiocarbon dated. The 14 bones show a large spread in strontium isotope ratios suggesting changes in landscape use or mobility over the lifetime of the individuals present in that grave. One bone fragment is particularly complex to interpret: bone 04227 (light blue in Figure 2). It is radiocarbon date matches with the more recent group of dates but its 87Sr/86Sr is lower (0.7111 compared to 0.7120, 0.7123 and 0.7127). It is, however, close to that of the bone fragment with the oldest date (0.7115—Figure 2). It could belong to that older individual but was impacted by an old-wood effect. Alternatively, it could belong to a third individual or diaphyseal turnover rates may be much more variable than initially thought. In any case, two distinct individuals could clearly be identified.
The case of Herstal 6 also shows that the combination of radiocarbon dates and 87Sr/86Sr enables to increase the MNI from two to three while, individually, each methodology only identified two (Figure 2). In Herstal 6, all 16 bone fragments but two have a very narrow range of 87Sr/86Sr (0.7115 to 0.7119—Figure 3). The two fragments with higher values (0.7129 and 0.7131) also have the more recent dates (Figure 2) and are present in different deposits (A and B). A similar case is seen in Grand Bois 41, where 11 out of 12 bone fragments have 87Sr/86Sr between 0.7109 and 0.7119 (Figure 3). The last bone fragment has a value of 0.7131 and is the bone fragment returning the oldest radiocarbon date.
The inclusion of the remains of multiple individuals in the same cremation might have different explanations; behavioural choices, simultaneous cremation events and/or using the same pyre site multiple times, whereby cremated remains from several ceremonies accumulated (un)intentionally (Booth and Brück Reference Booth and Brück2020). The presence of multiple individuals in one grave could result from people dying around the same time, cremated and interred together (Louwen Reference Louwen2021). However, the difference in radiocarbon dates suggest that the individuals were not burned at the same time. The large differences in 87Sr/86Sr further suggest in several of these cases that the individuals whose bones were found together consumed food from different areas and might actually have come from elsewhere. Seeing the large difference in calibrated radiocarbon dates (up to several centuries), one may wonder if any of this was intentional. The deceased could belong to socially related groups, emphasising certain social relations in death such as familial significance (Booth and Brück Reference Booth and Brück2020; Brück and Booth Reference Brück and Booth2022; Louwen Reference Louwen2021; Oestigaard and Goldhahn Reference Oestigaard and Goldhahn2006; Rebay-Salisbury Reference Rebay-Salisbury2018). Indeed, as family graves are found in modern cemeteries, they could have existed during the BA–IA to emphasise certain relations in death (Louwen Reference Louwen2021). Such multiple graves with possible family meaning were already observed in few Middle Bronze Age sites in Britain (Caswell and Roberts Reference Caswell and Roberts2018), in LBA-EAI urnfield cemeteries in southern Germany and Austria (Sørensen and Rebay-Salisbury Reference Sørensen and Rebay-Salisbury2023), in the IA in northern Europe (Ettel Reference Ettel2014) and were detected recently in some LBA–EIA sites in The Netherlands (Louwen Reference Louwen2021). Moreover, cremation enables the burying individuals from different generations together as the process transforms the corpse to small fragments of bones that can be kept for a certain amount of time (Louwen Reference Louwen2021). Nevertheless, whether the presence of several individuals within a cremation grave is intentional or not is not possible to assess with the current dataset and requires further investigation.
Conclusion
This study has clearly demonstrated that from a methodological point of view, though expensive, multi-sampling and combining osteological analyses with radiocarbon dates and 87Sr/86Sr on the same fragments of calcined bone can be used to better infer the MNI in cremation deposits, detecting multiple individuals in the same grave which is not possible to do using conventional osteological methods alone. It also shows that taking only one sample per grave might miss the complexity of the grave dynamics. It also highlights the crucial importance to carry out all analytical measurements on the same bone fragment to ensure that the results can be securely linked to the same individual. Further research is also needed to better understand how bone turnover might affect 87Sr/86Sr.
From an archaeological point of view, this research shows that the divergence of Herstal with the presence of double or even triple cremation burials is clearly not an isolated case but a more common phenomenon, during LBA–EIA in Belgium and probably other regions. Furthermore, it raises the question of intentionality linked to the presence of several individuals within a single cremation grave.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2024.82
Acknowledgments
This project was made possible thanks to the financial support of the FWO-F.R.S.-FNRS with the EOS project No. 30999782 CRUMBEL. Cremations, Urns and Mobility – Ancient Population Dynamics in Belgium. We would like to thank the Fonds Wetenschappelijk Onderzoek—Vlaanderen (FWO) for E. Stamataki and M. Hlad’s doctoral fellowships and the Fonds de la Recherche Scientifique (F.R.S.-FNRS) for A. Sengeløv’s doctoral fellowship. Nadine Mattielli, Wendy Debouge and Jeroen De Jong (G-TIME, ULB) are thanked for their help with the strontium isotope analyses. We acknowledge all the institutes that provided the cremated bone material from their collections: the Agentschap Onroerend Erfgoed, the Art & History Museum (RMAH), the Gallo-Roman Museum Tongeren, and the Herstal municipal museum. Vinciane Debaille thanks the FRS-FNRS for present support.
Open access
