The Power of Relics: The Curation of Human Bone in British Bronze Age Burials

In this article, the authors examine radiocarbon, histo-taphonomic, and contextual evidence for the deliberate curation, manipulation, and redeposition of human bone in British Bronze Age mortuary contexts. New radiocarbon dates and histological analyses are combined with existing data to explore the processes and practices that resulted in the incorporation of ‘relic’ fragments of bone in later graves, including evidence for the deliberate re-opening of previous burials and for funerary treatments such as excarnation and mummification. In some cases, fragments of human bone were curated outside the mortuary context. The authors consider what the treatment of human remains reveals about mortuary complexity in the Bronze Age, about relations between the living and the dead, and about attitudes to the body and concepts of the self.


INTRODUCTION
In recent years, detailed studies of grave goods from British Bronze Age burials have indicated that some objects were already old on deposition (e.g. Woodward & Hunter, 2015). Such items, it has been suggested, may have been heirlooms, retained and circulated to mark and maintain significant interpersonal and intergenerational relationships. The social and political implications of such practices have been the focus of some discussion, but the possibility that human remains from Bronze Age graves might also have been curated (that is, deliberately retained for years, decades, or even centuries after death to be deposited or redeposited in later mortuary contexts) has yet to be explored in detail. While archaeological investigation of British Bronze Age funerary practices has focused primarily on traditions of single burial, disarticulated, partially articulated, and fragmentary unburnt remains have been identified in many Chalcolithic and Early Bronze Age graves (Petersen, 1972;Gibson, 2004;Fowler, 2013: 108-69). Such finds have often been viewed as resulting from the accidental disturbance of earlier burials when graves were re-used for later interments. From c. 2100 BC onwards, cremation was common. Cremation burials frequently comprise only portions of the bodies of the dead, and sometimes include bones belonging to more than one individual (McKinley, 1997); taphonomic factors, or the death of several people around the same time, are usually invoked as explanations for the character of these deposits (McKinley, 1997: 142).
In this article, we explore the possibility that such deposits may have been the result of deliberate curation and redeposition of human bone. Recent work has identified a number of Bronze Age skeletons that show evidence of having previously been mummified (Booth et al., 2015). The best known is the mummy from Cladh Hallan on South Uist, found buried beneath the floor of a roundhouse of Late Bronze Age-Early Iron Age date (Parker Pearson et al., 2005). The practice was not limited to that period: formerly mummified remains have also been identified in several Chalcolithic and Early Bronze Age graves (Smith et al., 2016). The possibility that fragmentary, disarticulated, and burnt bones from Bronze Age mortuary contexts might likewise have been preserved and curated therefore deserves examination.
The research presented here builds on a previous article reporting the results of chronological modelling of 189 radiocarbon dates obtained from possibly curated bone from British Bronze Age burials and settlements (Booth & Brück, 2020). Here, we focus on dates from mortuary contexts alone (finds from settlements will be treated in a separate article) to consider whether there is consistent evidence for the deliberate curation of human bone in Bronze Age burials. As detailed consideration of individual depositional contexts did not form part of our 2020 article, here we combine the dating evidence with contextual and histological (bone microstructure) analyses to examine the variety of processes and practices that facilitated the creation of such 'relics'. We consider what the curation of human bone in British Bronze Age burials can tell us about attitudes to the body and concepts of the self, about the treatment of the bodies of the dead, and about the role of the dead in the world of the living.
This study adds to existing work on the interpretation of fragmentary human remains from mortuary and non-mortuary contexts, such as discussion of the composition, taphonomy, and context of fragmentary human remains from Neolithic monuments, as well as from settlement sites of Neolithic, Bronze Age, and Iron Age date in Britain and continental Europe (e.g. Stapel, 1999;Smith & Brickley, 2009;Müller-Scheeßel, 2013). Both within and beyond European prehistory, attention has been directed at the complex, variable, and often protracted nature of funerary practice, including multi-stage burial rites that required either the preservation or structured dissolution of the body (e.g. Rebay- Salisbury et al., 2010;Weiss-Krejci, 2010;Gramsch, 2013). Practices that involved the deliberate re-opening of existing burials for a variety of purposes have been addressed (e.g. Aspöck et al., 2020), as has the role of human remains in non-funerary settings, including their deliberate curation in more recent historical contexts (e.g. Weiss-Krejci, 2005). The present study is, however, the first to employ radiocarbon dating in a sustained way to identify curation; it also adds to the small body of research to date that has combined histological analysis with other evidence to understand complex mortuary practices in European prehistory (e.g. Booth & Madgwick, 2016).
Of course, the terms 'relic' and 'heirloom' have a specific history and origin, and it could be argued that they are irrelevant in understanding Bronze Age practices. However, the social role of heirlooms and relics in recent European history provides interesting points of comparison and contrast. Medieval relics, for example, were embodiments of the sacred. They channelled supernatural and apotropaic powers and were thought to be social agents with their own life force (e.g. Räsänen et al., 2016). They conveyed authority on those who owned or had access to them, and legitimated earthly as well as spiritual power. In more recent years, the bodies of the dead-the disappeared of the Spanish Civil War, for example-have often acted as particular foci of political struggles (e.g. Verdery, 1999). In other contemporary contexts, heirlooms perpetuate the presence of the dead among the living and are bound up with important aspects of social identity (Weiner, 1992). The affective power of relics and heirlooms is generated through sensory and aesthetic engagement, so that the performances and practices in which they are enmeshed are central to understanding their agency. Thus, although we cannot assume that curated objects were viewed in the same way in the Bronze Age, historical and anthropological studies of heirlooms and relics usefully illuminate the range of significant social roles that such items often play.

SAMPLING AND METHODOLOGY
In total, we collated eighty-one radiocarbon dates from fifteen graves on thirteen different mortuary sites ( Figure S1 and Table S1 in Supplementary Material); eleven of these thirteen sites are Chalcolithic or Early Bronze Age in date, and hence most of the observations here apply to these periods. Sixty-four existing dates were used, and we generated seventeen new dates for the purposes of this project. Eleven of the graves studied contained only unburnt bone; the other four yielded deposits of burnt bone belonging to more than one individual. In order to test whether the unburnt bone was significantly older than its depositional context, we compared the dates of potentially curated human remains with the dates of associated articulated burials (in one instance an associated unburnt animal bone provided the proposed date of deposition). For the cremation graves, we compared the dates of burnt bone from different individuals found commingled in the same context. Modern osteological analyses were available for all but two graves, ensuring that the same individuals were not sampled twice (for the two exceptions, contextual information was sufficient to be confident of this). Except for one grave, all were modern excavations for which detailed contextual information was available; this was crucial to understanding taphonomic and depositional processes and also ensured that the association between dated samples was secure. In two instances, because no associated short-life material was available, we compared the dates of human bone with date ranges based on the typology of accompanying artefacts.
Bones that are older than their depositional context may, of course, have become accidentally incorporated into a later context. Our previous article (Booth & Brück, 2020) employed chronological modelling to demonstrate that a higherthan-expected proportion of samples from both mortuary and settlement sites were anomalously old and that there was a consistent pattern of around two generations in the interval between the date of death and the date of deposition, suggesting a deliberate cultural practice rather than accidental incorporation (the methodology employed to model the dates is summarized in the supplementary information). Our chronological modelling indicated that, even where individual samples could have been curated for considerable periods, the bones were likely to have been curated for decades rather than centuries. Comparison with a control sample suggested that bones which were not anomalously old may also have been curated, but for short periods that were hard to discern within the range of radiocarbon error. We argued that the marine reservoir effect is unlikely to have affected the dates we collated, as stable isotope analysis of human remains from Bronze Age Britain indicates very low levels of consumption of marine or freshwater resources (Parker Pearson et al., 2019).
Histological analyses, either already published or newly conducted by us, were undertaken on some dated bones to investigate variability in post-mortem treatment (Table S1). Previous studies indicate that bodies buried intact and soon after death in generally dry, aerobic environments will suffer high levels of bacterial bioerosion to the internal bone microstructures visible through micro-CT analysis (Jans et al., 2004;Booth, 2016;. By contrast, low levels of bacterial bioerosion in bones from the same kind of burial environment correlate with forms of postmortem treatment that inhibit bodily decomposition, such as mummification, or that rapidly remove soft tissue, such as excarnation and dismemberment (Booth et al., 2015;Booth, 2016;Brönnimann et al., 2018).

DEPOSITIONAL PROCESS AND MORTUARY PRACTICE
At seven of the thirteen sites considered, chronological modelling indicated that samples were anomalously old (Table S1). Although these bones may have been accidentally incorporated into later graves, the evidence points to a range of complex mortuary practices. The histological analyses of bone from graves dated mostly to the Chalcolithic or Early Bronze Age all demonstrate very low levels of bacterial bioerosion to the internal microstructures, suggesting post-mortem processes resulting in arrested microbial attack, such as excarnation or mummification. At Melton Quarry in East Yorkshire, for example, the disarticulated and incomplete remains of an infant were nestled between the torso and legs of a complete inhumation burial (Fraser Brown and Lauren McIntyre, Oxford Archaeology South, pers. comm. 2018). Comparison of radiocarbon dates for these two individuals indicated that the bones of the infant were between 189 and 348 years older than the articulated burial (at 68 per cent confidence; note that throughout this article, 68 per cent confidence intervals are quoted in the text to provide a better sense of probable 'true' values; 95 per cent confidence intervals are, however, also provided in the figures) ( Figure 1). Histological analysis indicated that the infant bones were subject to low levels of bacterial attack to the internal bone microstructure, contrasting with a bone sample from the articulated skeleton, which showed extensive bacterial bioerosion. The deposition of the infant bones in a disarticulated state suggests that its body had been excarnated immediately after death. Assemblages of items that were probably deposited in organic bags are a common feature of Chalcolithic and Early Bronze Age burials (Cooper et al., 2019), and it is possible that these infant bones were kept in such a container, perhaps even worn on the body of the articulated individual during life. The complex afterlives of the dead are also illuminated by a slab-lined grave at Cnip Headland on the Isle of Lewis (Lelong et al., 2018; Figure 2). It contained the incomplete and partially articulated skeleton of a probably male adolescent as well as the disarticulated remains of at least two other adults. The adolescent had been laid on his right side and the excavators suggest that the body was buried at an advanced stage of decomposition. Several articulated thoracic vertebrae were out of alignment with the rest of the spine, and the left fibula and right humerus were also dislocated, indicating that these elements may have been skeletonized at the time of burial. A significant gap between the head and torso, and between the torso and lower body, implies that the body may have been buried in several parts, while the presence of most of the left hand and wrist bones suggests that the body was partly fleshed on burial; many other bones were missing, including most of the right arm and both feet. Histological analysis indicates very low Figure 1. a) Chronological model of the radiocarbon dates from Melton Quarry burials 1008 and 1009. b) Probability distribution of the difference between the radiocarbon dates from the articulated inhumation burial (SUERC 72661) and the disarticulated infant bones (SUERC 72666). Note that in this figure and all other probability distribution figures in this article, light grey represents the unmodelled difference, comparing the dates directly with no assumptions. Dark grey represents the modelled difference (posterior density estimate), comparing the dates within the constraints applied by assumed or observed relationships with other dated materials included in the model. Quoted differences reflect modelled differences. levels of bacterial bioerosion. It is possible that the adolescent had been excarnated and then buried before complete skeletonization. However, a disarticulated metacarpal recovered from the base of the grave in a layer of sand stratigraphically predating the deposition of the adolescent returned a date that was anomalously young, i.e. postdating the age of death of the adolescent. This suggests that the bones of the adolescent were already old when they were interred in the grave. Calculation of  Lelong et al., 2018: illus. 5). b) Chronological model of the radiocarbon dates from Cnip Headland. c) Probability distribution of the difference between the radiocarbon dates from the partially articulated adolescent male (Sk1) and the disarticulated metacarpal from the base of the grave (SUERC 30860). SUERC 30859 and 39854 are dates for other disarticulated bones from the same context as Sk1 but belonging to other individuals. Figure 2a reproduced by permission of Ingrid Shearer and Olivia Lelong. modelled differences in OxCal 4.4 suggests that the bones of the adolescent had been curated for between three and eighty-two years (at 68 per cent confidence) prior to deposition (Figure 2). The partial articulation of these remains and the low level of bacterial attack to the bones thus appear more consistent with a form of post-mortem treatment such as mummification than with excarnation.
It is occasionally possible to infer where old human remains were stored before they were deposited in their excavated context. At Windmill Fields, Ingleby Barwick, Stockton-upon-Tees in North Yorkshire, the contracted inhumation burial of an adult female was accompanied by disarticulated crania and longbones representing at least three individuals: a possibly adolescent female, an adult male, and an adult female (Annis et al., 1997; Figure 3). The disarticulated remains had been neatly stacked just in front of the body of the female. The two adult crania are 59-179 years older (at 68 per cent confidence) than the articulated burial ( Figure 4). Dark staining in a nearby pit suggests it originally contained a wooden coffin or mortuary structure. This feature yielded the partial and disarticulated remains of two adult males, as well as several other skull fragments. Histological analysis of one of the disarticulated adult longbones from the wooden structure revealed little bioerosion, consistent with the body having been excarnated (Booth et al., 2015(Booth et al., : 1167. This contrasted with the femur from the female inhumation burial, whose abundant bacterial tunnelling suggests that it was a primary burial (i.e. a complete body buried in a fresh state and left in its original place of deposition). The two adult crania that accompanied the female inhumation are contemporary with the deposit of disarticulated remains in the wooden structure (Booth et al., 2015(Booth et al., : 1197. It seems reasonable to speculate in this instance that the bones placed in front of the body of the female had been retrieved from the nearby wooden coffin or mortuary structure. It has long been accepted that primary burials were frequently re-opened (Petersen, 1972), sometimes to insert another body, but probably also to encounter and retrieve the bones of the dead (Gibson, 2013) or grave goods. At South Dumpton Down, in Kent, a grave contained two complete articulated inhumation burials of Early Bronze Age date (Perkins, 1995). One of these was accompanied by the disarticulated mandible of a third individual. We radiocarbon-dated that mandible and the associated inhumation burial, but the mandible was not anomalously old. Nonetheless, it derived from a body that had already decomposed. The mandible was not one of the bones analysed histologically and hence we do not know whether it was retrieved from a primary burial or from excarnated remains. However, there is good evidence for the revisiting of graves at this site. A deep shaft grave nearby held the remains of five individuals who had been deposited in sequence. Several of their skulls were missing, suggesting that each time the grave was  opened to receive another burial, selected bones from previous interments were removed for re-use and redeposition elsewhere. It is therefore possible that the mandible from the double inhumation was retrieved from the shaft grave for deliberate redeposition. If the exhumation and redeposition of such bones took place several months or years rather than decades or centuries after their original burial, they are unlikely to show up as anomalously old using the tests provided in OxCal 4.4.
Elsewhere, bones may have been curated among the living. A femur fragment from a pit at Cotswold Community near Ashton Keynes in Wiltshire (Smith et al., 2010) was 5-175 years older (at 68 per cent confidence) than burnt animal bone and ash charcoal from the same context ( Figure 5). The pit also yielded a second femur fragment, probably from the same individual, alongside sherds of Beaker pottery, charcoal, burnt stone, and charred plant remains. Isolated pits and pit clusters have frequently been interpreted as evidence of Chalcolithic and Early Bronze Age settlement (Brück, 1999), for their contents often include the residues of domestic activity. It has been argued that this material may have been collected and deposited to mark significant places or commemorate important people and events (Garrow, 2007). The series of scattered pits identified at Cotswold Community may represent seasonal or annual visits to this location over many centuries. Among communities that were relatively mobile, carrying portions of the significant dead may have been a means of demonstrating belonging. So too the deposition of curated remains might have worked to maintain links between people and place.
The evidence from sites such as South Dumpton Down suggests that specific body parts may have been chosen for curation and/or redeposition. Although the sample was small, skull and longbone fragments were the most commonly curated skeletal elements in graves containing unburnt bone (supplementary material: Table S1). This observation should, however, be treated with caution, as we did not re-examine the original osteological collections. Modern osteoarchaeological analyses were available in most cases, but skeletal part representation was not a particular consideration in those studies, and critical assessment of the representation of different skeletal elements was therefore not possible.

THE CURATION OF BURNT BONE
Because of the potential impact of the old wood effect on cremated bone (Olsen et al., 2013;Snoeck et al., 2014), just four burials containing possibly curated cremated remains were included in our study. We did not date cremated bone accompanying inhumation burials because the dates obtained from different sample types might not be comparable. Two of the cremation burials we analysed produced anomalously early dates, and the offset between date of death and date of deposition for these samples was no greater than for unburnt bones that produced significantly anomalous dates. While we cannot rule out the possibility that the old wood effect could be responsible for these cremated bones returning dates that appear too old, a scenario where cremated bones had been curated is a plausible alternative.
At Trelowthas in Cornwall, a stone cist had been filled with the cremated bones of multiple individuals (J. Nowakowski, pers. comm. 2018). An urn holding the cremated remains of at least two further individuals was then inserted into this bone deposit. The remains in the urn were of individuals 3-72 years older than the bone from the deposit into which the urn was inserted ( Figure 6). We suggest that the curated bones were stored in the urn before they were deposited in the ground. During that time, they may have been kept outside the mortuary context, in a location where they could be encountered and handled by the living. Other recent work in the same region suggests that mortuary sites were often foci for repeated Figure 5. a) Chronological model of the radiocarbon dates from Cotswold Community. b) Probability distribution of the difference between the radiocarbon dates from the femur fragment (BRAMS 1448) and the ash charcoal (BRAMS 1281) and burnt animal bone (BRAMS 1283) (combined). Figure 6. a) Chronological model of the radiocarbon dates from Trelowthas. b and c) Probability distribution of the difference between the radiocarbon date from a burnt bone in the urn (BRAMS 1290) and a modelled likely end date for the deposit into which the urn was placed (lower spit 3 end) based on dates from two samples of burnt bone from the latter deposit (BRAMS 1288 and 1289). c) Probability distribution of the difference between the radiocarbon date from a burnt bone in the urn (BRAMS 1290) and burnt bone from the cist into which this vessel was deposited (BRAMS 1289). BRAMS 1292 is from burnt bone found beneath the basal stone of the cist.
visits that involved the deposition of small quantities of bone, pottery, charcoal, and other materials (Frieman & Lewis, 2021).
The pit at the centre of a ring-ditch at Whitton Hill in Northumberland yielded 21.6 kg of burnt bone, representing the remains of at least twenty-four individuals, including seven or eight children and sixteen adults or adolescents (Gamble & Fowler, 2013). The only individual who could be sexed was female. The bone was found in a single deposit in the pit's upper fill. There was no indication that the pit had been revisited, implying that the cremated remains were deposited in a single event. Three cremated human bones from discrete individuals had radiocarbon dates. Two of these bones were 3-115 and 3-37 years older, respectively, than the third (Figure 7). This may indicate re-use of the pyre site so that fragments of bone from previous cremations were accidentally incorporated into later burials, but, in the context of the other evidence presented here, it is equally plausible that the modelled dates indicate deliberate curation of burnt bone, particularly given the very large number of individuals represented in this deposit.

RELATIONS WITH THE DEAD
The evidence presented indicates that fragmentary, partially articulated, and disarticulated bones from Bronze Age mortuary contexts were often deliberately curated. Although some are likely to have been retrieved from primary burials, practices such as excarnation, mummification, and cremation facilitated curation outside the mortuary context before subsequent redeposition. While some finds may be the result of accidental incorporation of bone fragments, for example when a grave was re-used, the curation of the bones of the significant dead is well-attested ethnographically, as several archaeologists have discussed (e.g. Weiss-Krejci, 2011;Armit, 2017). The Ongee of the Andaman Islands, for example, make human bones into body ornaments to commemorate interpersonal relationships and to ensure the continued efficacy of the ancestral dead in the world of the living (Pandya, 2009: 124-25).
In order to understand why the dead were curated, it is necessary to consider who might have been subject to these practices. The similarities in the treatment of the bones discussed here and contemporary heirloom artefacts (e.g. Woodward & Hunter, 2015) suggest that fragmentation of the body was viewed as a productive means of creating and maintaining relational identities. We may infer from the careful positioning of curated bone relative to the bodies of subsequent burials at the sites described here that links between the living and the dead formed a significant component of social identities in the Chalcolithic and Early Bronze Age, and that these were viewed as intimate bodily ties. It has often been argued that the spatial relationships between burials in Early Bronze Age barrows or between different barrows within a cemetery reflect family relationships (Garwood, 1991;Mizoguchi, 1992). Set within this context, it seems likely that curated bones from mortuary contexts were considered to belong to kin, although the possibility that they gave material form to other kinds of social bonds cannot be excluded.
Possible kinship links may be supported by recent aDNA analysis of samples from one of the graves included in this study. The well-known multiple burial known as the Boscombe Bowmen at Boscombe Down in Wiltshire contained the remains of several adults and children (Fitzpatrick, 2011;McKinley, 2011;Figure 8). At the base of the grave were the inhumation burials of a 5-6-year-old child and an adult male (burial 25004). Beneath the adult male was a bundle of disarticulated bones comprising selected skeletal elements from four other individuals (two adult males, one subadult male, and a juvenile; bone deposit 25008), predominantly longbone fragments from the left side of the body. Two crania and part of a mandible lying at the feet of burial 25004 may have belonged to the two adults represented in the bone bundle.
Genetic analysis of burial 25004 and one of the crania (ON10, an adult male aged 25-30) indicated that the disarticulated skull came from a second to third degree relative of 25004 (Olalde et al., 2018;Booth et al., 2021). Burial 25004 and ON10 belong to different maternal lineages but the same paternal lineage: the skull may have belonged to a genetic paternal cousin, half-sibling, (great) uncle, or (great) grandfather of the articulated male. This suggests that disarticulated remains at least sometimes belonged to genetic relatives of the primary burials with which they were deposited. Although ON10 was not radiocarbon-dated, three disarticulated femora from the grave (including one from bone bundle 25008) were probably significantly older than articulated burial 25004 (Figure 9).
The strontium stable isotope analysis of enamel in teeth from the adult male 25004 and the two crania found at his feet suggests that they all undertook similar journeys in childhood (Evans et al., 2006). If the two individuals represented by the skulls were broadly contemporary with burial 25004, all three could have undertaken this journey together in life. Alternatively, if the skulls were curated, they may have been transported to the burial site many years after death. If that were the case, then the decision to curate and redeposit the two crania alongside an articulated body who had a similar life history could imply that the identities of the individuals to whom the skulls belonged were known and remembered. Altogether, the evidence suggests that biography as well as kinship formed part of the decision to curate and redeposit human bone in particular ways. Curated remains may have belonged to specific kin who were remembered as individuals rather than as part of a generalized ancestry. This argument is supported by calculation of the median period of curation for the burials discussed. Calculation of the differences between date of death and date of deposition by combining interval probability distributions generated in the BChron software in R indicates a median period of curation of ninety-five years with an interquartile range of 148 ( Figure 10; first quartile = 172, third quartile = 24; Haslett & Parnell, 2008;R Core Team, 2013; see Booth & Brück, 2020 for details Figure 9. a) Chronological model of the radiocarbon dates from the Boscombe Bowmen primary deposit built in OxCal. b) Probability distribution of the difference between the radiocarbon dates from a disarticulated femur of an adult male from bone deposit 25008 (OxA 13543) and the accompanying articulated burial (25004; OxA 13624). c) Probability distribution of the difference between the radiocarbon dates from a disarticulated femur of an adult male from context 25010 (OxA 13542) and the articulated burial (25004). d) Probability distribution of the difference between the radiocarbon dates from a disarticulated femur of a subadult from context 25010 (OxA 13681) and the articulated burial (25004). The remaining dates are from other burials from the same grave: the cremation burial of an infant (OxA 13972) and the articulated inhumation burials of two juveniles (OxA 13598 and 13599). on statistical methods). Assmann (2008) makes a useful distinction between communicative and cultural memory: the former can be characterized as living memory, transmitted in the context of everyday interpersonal interaction, while the latter refers to processes of institutionalization by which memories are given an external material form, for example in myths, symbols, and rituals. Communicative memory, Assmann argues, can be maintained across three to four generations or 80-100 years. This correlates neatly with the median period of curation for the human remains discussed here and suggests that the identities of those whose remains were curated in Bronze Age mortuary contexts were known to those who placed them in the grave.
Who was selected for curation? Four males and two females (including probable and possible identifications) could be identified when considering only contexts with positive radiocarbon evidence for curation and curated remains whose sex could be determined (supplementary material: Table S1). In other words, individuals of any gender could be viewed as significant ancestors. Age category estimates were available for curated bone from six sites. Four were adults, one was an infant aged 2-4 months, one was an adolescent aged 12-20, and one was a subadult of undetermined age (supplementary material: Table S1). Although adolescents may have been considered to be social adults, the curation of children and young people is noteworthy and suggests that Bronze Age kinship structures, ideologies, and determinations of status were quite different from our own.
Kinship is not biologically determined (Brück, 2021), and other types of social relations may have been equally important in determining whose bones were curated and how these were treated. At Wilsford in Wiltshire, a bone whistle made from a human femur accompanied the inhumation of an adult male (Woodward & Needham, 2012;Woodward & Hunter, 2015: 114) with other grave goods suggesting a date range of 1950-1700 BC. Several possible shamans' graves have been identified in this period (Piggott, 1962: 96), and this too may be the burial of a ritual specialist. We sampled the whistle for radiocarbon dating and compared it to the inferred date range (the inhumation itself was reinterred in the nineteenth century). The result indicated that the bone whistle was unlikely to have been very old on deposition, suggesting that the person to whom the femur originally belonged may have been known to those who deposited it in the grave. Although kinship may have been a factor determining selection, it is worth considering other scenarios too. A flute made from the bone of an esteemed elder, a ritual specialist belonging to a previous generation, or even an enemy might have been invested with particular power, although evidence to support one interpretation over another is not currently available.
Certain Bronze Age artefacts, such as those made of raw materials with unusual physical properties (e.g. jet and amber), may have been considered to be active social agents with their own spirit or soul (Jones, 2001;Sheridan & Shortland, 2003). We suggest that objects made from human bone, such as the whistle from Wilsford, were viewed as especially powerful. The socio-political power that may be derived from the ability to access and display the bones of the significant dead has been a major focus of discussion in archaeology and anthropology (e.g. Verdery, 1999;Weiss-Krejci, 2011). In Bronze Age Britain, the deposition of curated bone belonging to known kin and significant others in a mortuary context may have acted as a means of expressing key socio-political relations. The bodily intimacy between deceased individuals, however, also reminds us that these were often affective relationships: in Chalcolithic and Early Bronze Age Britain, emotional attachments may have been just as important as political exigencies in shaping funerary practice (see Tarlow, 2000). Moreover, the final deposition of such relics at the point where individuals were disappearing from living memory suggests that these might equally have been acts of forgetting, designed to allow the reorientation of social ties.

CONCLUSION
Our research indicates that fragmentary, disarticulated, and partial human remains from graves (both burnt and unburnt) were sometimes significantly older than their depositional context, while elsewhere bone was redeposited after a relatively short period. In some cases, bone had been curated, probably outside the mortuary domain. The deposition of bones belonging to long-dead individuals calls into question the assumption that grave goods are a direct reflection of social identity, for the identities of the curated dead may have been mythologized and reconfigured over the years. The curation of human bone was facilitated by a variety of post-mortem practices, including excarnation, cremation, and mummification. It seems possible that curated bone was viewed as a powerful, animate material allowing the dead to play a significant social role in the lives of the living.
Skeletal elements were also exhumed from existing burials for redeposition with other bodies. Bronze Age people were thus used to sensuous engagement with human remains-whether fleshed, partially decomposed, fully skeletonized, or cremated-and it is possible that curated remains were also carried on the bodies of the living. Practices such as excarnation and cremation aided the fragmentation of the body, and it is evident that the disaggregation of bodily elements was not considered ontologically problematic. In the Bronze Age, the fragmentation of the body made the combination and recombination of bodily elements possible. This suggests that identity was defined in relational terms: the person was composed of multiple elements brought together, disaggregated, and reordered through different forms of social practice including mortuary rituals (Brück, 2019). This calls into question the concepts of the self-as a defined, enduring locus of power-that underpin dominant narratives of the Bronze Age in which competition between warrior chiefs resulted in the development of stratified societies (Brück & Fontijn, 2013).
The physical incorporation of parts of different bodies into the same mortuary deposits indicates they were regarded as closely related, and the relatively short temporal interval between death and deposition in many cases points to the curated remains being the bones of known and significant ancestors, i.e. people whose particular biographies could be recounted by their immediate descendants. The curation of the bones of children suggests that the idea of ancestry may, however, not have been based solely on lineal descent. In other cases, several generations had probably elapsed, and the original identity of the ancestral bone may not have been known. This is not to assume that kinship was based solely on biogenetic links: bones exhumed during later re-use of a barrow may have been identified as those of ancestors, regardless of blood ties. Of course, interpersonal ties other than kinship may also have generated the sorts of close, affective bonds suggested by practices of curation. Claiming links with the dead doubtless constituted an effective way of expressing social and political affinities, defining concepts of belonging and exclusion, and negotiating rights to significant roles and resources. Funerary rites must, however, not only be construed as a means of reproducing socio-political power: the intimate bodily connections between those who occupied the same grave indicates that emotional bonds were also significant factors in the decision to curate the remains of the Bronze Age dead. Woodward, A. & Needham, S. 2012 Le pouvoir des reliques: la conservation intentionnelle d'ossements humains dans les sépultures de l'âge du Bronze des Îles britanniques Les auteurs de cet article présentent les données radiocarbone, taphonomiques, histologiques et contextuelles concernant la conservation intentionnelle, la manipulation et la redéposition d'ossements humains provenant de sépultures de l'âge du Bronze fouillées dans les Îles britanniques. En combinant des éléments publiés avec de nouvelles dates radiocarbone et analyses histologiques, ils examinent les processus et pratiques qui ont mené à l'intégration d'ossements « reliques » en contextes funéraires, y compris la réouverture de sépultures antérieures et les divers traitements des dépouilles tels que l'excarnation et la momification. Dans certains cas, on a conservé des ossements humains en dehors des lieux de sépulture. Les auteurs considèrent ce que le traitement des cadavres révèle sur la complexité des pratiques funéraires de l'âge du Bronze, sur les rapports entre les vivants et les morts, sur les attitudes envers le corps et sur le concept de soi. Translation by Madeleine Hummler