Reassessing Community Cemeteries: Cremation Burials in Britain during the Middle Bronze Age (c. 1600–1150 cal bc)

The Middle Bronze Age (c. 1600–1150 cal bc) in Britain is traditionally understood to represent a major funerary transition. This is a transformation from a heterogeneous funerary rite, largely encompassing inhumations and cremations in burial mounds and often accompanied by grave goods, to a homogeneous and unadorned cremation-based practice. Despite a huge expansion in the number of well excavated, radiocarbon dated, and osteologically analysed sites in the last three decades, current interpretations of Middle Bronze Age cremation burials still rely upon a seminal paper by Ellison (1980), which proposed that they comprise and represent an entire community. This paper analyses 378 cremation sites containing at least 3133 burials which represent all those that can be confidently dated to the Middle Bronze Age in Britain. The new analysis demonstrates that relatively few sites can be characterised as community cemeteries and that there are substantially more contemporary settlement sites, though few contemporary settlements are in close proximity to the cemeteries. The identifiable characteristics of cremation-based funerary practices are consistent across Britain with little evidence for social differentiation at the point of burial. It is also evident that only a minority of the population received a cremation burial. There is a substantial decrease in archaeologically visible funerary activity from the preceding Early Bronze Age (c. 2200–1600 cal bc) and a further decrease in the proceeding Late Bronze Age (c. 1150–800 cal bc) in Britain. This is comparable in form, and partially in sequence, to Bronze Age funerary practices in Ireland and several regions in North-west Europe. RÉSUMÉ Réévaluation des cimetières collectifs: sépultures à incinération en Grande-Bretagne durant l’âge du bronze moyen (env. 1600–1150 cal av. J.-C.), de Edward Caswell & Benjamin W. Roberts L’âge du bronze moyen (env. 1600–1150 cal av. J.-C. en Grande-Bretagne est traditionnellement considéré comme représentant une transition funéraire majeure. Celle-ci consiste en une transformation d’un rite funéraire hétérogène incluant essentiellement inhumations et incinérations dans des tertres funéraires, souvent accompagnées de mobilier funéraire, à une pratique homogène et sans mobilier reposant sur des incinérations.Malgré une vaste expansion du nombre de sites soigneusement fouillés, datés au C14 et analysés ostéologiquement au cours des trois dernières décennies, les interprétations actuelles des sépultures à incinération de l’âge du bronze moyen s’appuient toujours sur un article élémentaire d’Ellison (1980). Alternativement,le présent article analyse 378 sites d’incinération contenant au moins 3133 sépultures qui représentent toutes celles que l’on peut dater avec certitude de l’âge du bronze moyen en Grande-Bretagne. La nouvelle analyse démontre que relativement peu de sites peuvent être caractérisés comme cimetières collectifs et qu’il y a substantiellement plus de sites d’occupation contemporains, bien que peu très proches des cimetières collectifs. Les caractéristiques identifiables des pratiques funéraires reposant sur l’incinération sont cmpatibles à travers toute la Grande-Bretagne avec peu de preuves de différentiation sociale au moment de l’inhumation. Il est également évident que seule une minorité de la population bénéficiait d’une inhumation à incinération. Il y a une diminution substantielle dans les activités funéraires visibles à l’archéologie du précédent âge du bronze ancien (env. 2200–1600 cal av. J.-C.) et une diminution encore plus importante à l’âge du bronze final qui a suivi (env. 1150–800 cal av. J.-C.) en Grande-Bretagne, ce qui est comparable dans sa forme et partiellement dans sa séquence , aux pratiques funéraires de l’âge du bronze en Irlande et dans plusieurs régions du nord-ouest de l’Europe. ZUSSAMENFASSUNG Eine Neubewertung von gemeinschaftlichen Gräberfeldern: Brandbestattungen in Großbritannien während der Mittelbronzezeit (ca. 1600–1150 cal bc), von Edward Caswell und Benjamin W. Roberts Die Mittelbronzezeit (ca. 1600–1150 cal bc) in Großbritannien wird traditionell als Zeit eines wichtigen Übergangs in der Bestattungssitte gesehen. Dieser Wandel vollzieht sich von einem heterogenen Begräbnisritual in Grabhügeln, in denen sich Körper- und Brandbestattungen, häuftig mit Beigaben, befinden, zu einer homogenen Begräbnisart bei der beigabenlose Brandbesttatungen vorherrschen. Trotz immenser Zunahme der Zahl gut ausgegrabener, C14-datierter und anthropologisch untersuchter Fundstellen in den letzten drei Jahrzehnten, beziehen sich derzeitigen Interpretationen mittelbronzezeitlicher Brandbestattungen noch immer auf den wegweisenden Aufsatz von Ellison (1980). Der hier vorliegende Beitrag untersucht 378 Brandbestattungsplätze mit insgesamt mindestens 3133 Beisetzungen, bei denen die Datierung in die britische Mittelbronzezeit sicher ist. Die neue Untersuchung zeigt, dass nur recht wenige Fundplätze als Gräberfelder einer Gemeinschaft charakterisiert werden können, und dass es deutlich mehr zeitgleiche Siedlungen gibt, wenn auch nur wenige in unmittelbarer Nähe zu den Gräberfeldern. Die erkennbaren Charakteristiken von auf diesen Brandbestattungen sind in ganz Großbritannien einheitlich und liefern nur wenigen Hinweisen auf eine soziale Differenzierung während der Bestattung. Es wird auch deutlich, dass nur eine Minderheit der Bevölkerung eine Brandbestattung erfuhr. In Großbritannien ist ab der vorhergehenden Frühbronzezeit (ca. 2200–1600 cal bc) ein erheblicher Rückgang archäologisch sichtbarer Begräbnisaktivitäten feststellbar und ein weiterer Rückgang in der nachfolgenden Spätbronzezeit (1150–800 cal bc). Dies ist in Form und teilweise im Ablauf vergleichbar mit den bronzezeitlichen Begräbnispraktiken in Irland und in mehreren Regionen Nordwesteuropas. RESUMEN Reevaluando los cementerios: los enterramientos de cremación durante el Bronce Medio (ca. 1600–1150 cal bc), por Edward Caswell y Benjamin W. Roberts El Bronce Medio (ca. 1600–1150 cal bc) en Gran Bretaña se ha entendido tradicionalmente como una importante transición en las prácticas funerarias. Esto implica la transformación de un ritual funerario heterogéneo, unido en gran medida a las inhumaciones y cremaciones en túmulos y generalmente acompañado de ajuares, a una práctica funeraria homogénea y sencilla basada en la cremación. A pesar del aumento en el número de sitios arqueológicos bien excavados, datados y del número de análisis osteológicos en las últimas tres décadas, las interpretaciones actuales de los depósitos funerarios del Bronce Medio aún se basan en el influyente artículo de Ellison (1980). Como alternativa, este artículo analiza 378 cremaciones que contienen, al menos, 3133 enterramientos entre los que se incluyen todos aquéllos que pueden datarse con seguridad en el Bronce Medio en Gran Bretaña. Este nuevo análisis demuestra que relativamente pocos sitios pueden ser caracterizados como cementerios y que hay un número sustancial de asentamientos contemporáneos, algunos de ellos en las proximidades de los cementerios. Las características identificables de estas prácticas funerarias por cremación son consistentes a lo largo de Gran Bretaña con pocas evidencias de diferenciación social en el momento de enterramiento. Es igualmente evidente que solo una minoría de la población recibe un enterramiento por cremación. Existe un decrecimiento sustancial en la visibilidad arqueológica de la actividad funeraria desde el precedente Bronce Inicial (ca. 2200–1600 cal bc) y un mayor decrecimiento en el posterior Bronce Final (ca. 1150–800 cal bc) en Gran Bretaña. Esto es comparable en forma, y parcialmente en secuencia, con las prácticas funerarias en Irlanda y en algunas regiones del noroeste de Europa.

The Middle Bronze Age (c.1600-1150 cal BC) in Britain is characterised in broad archaeological narratives by the major expansion of settlements and bronze and gold metalwork hoards throughout Britain, as well as the construction of field systems in central and southern England (Darvill 1996, 108-32;Bradley 2007, ch. 4;Yates 2007;Cunliffe 2013, 266-7).Contemporaneously, the funerary record is overwhelmingly dominated by cremation burials.The presence of certain Middle Bronze Age cremation sites such as Itford Hill, Sussex (Holden 1972);Stansted Airport, Essex (Cooke et al. 2008); Harehope, Peebleshire (Jobey 1980); and Black Patch, East Sussex (Drewett 1982) in close proximity to contemporary settlements has encouraged interpretations that they contain entire communities from these nearby settlements (eg, Bradley 1981;2007, 185;Darvill 1996, 116-17;2010, 222).The interpretation that Middle Bronze Age funerary sites throughout Britain are representative of community cemeteriesgroups of cremation burials that are associated with a specific community and contain all members of that communitycan be traced back to a seminal paper by Ellison (1980).However, despite the immense increase in the excavation and radiocarbon dating of Middle Bronze Age cremation burials over the last 35 years, and the refinement and widespread application of osteological analyses on cremated bone, there has been no reappraisal of the community cemetery interpretation.
Similarly, there has never been an analysis of Middle Bronze Age cremation burials that goes beyond a region larger than southern Britain.
In the last few years, there has been a substantial advance in the analysis and interpretation of cremated human remains.In particular, there has been an increasing emphasis on: the standardisation of recording (McKinley 2004); the development of new analytical and experimental methods (eg, Marshall 2011;Depierre 2014;Kuijt et al. 2014;Snoeck et al. 2015;Thompson 2015); as well as an understanding of the highly varied chaîne opératoires in the creation of superficially similar cremation burials (eg, Rebay-Salisbury 2010; Marshall 2011; Appleby 2013; Depierre 2014).There is also now a broader recognition of the importance of analysing this funerary phenomenon which has been far too frequently overlooked in broader mortuary studies due to it being less visible in the archaeological record (see Rebay-Salisbury 2010;Kuijt et al. 2014;Barceló et al. 2014;Capuzzo & Barceló 2015;Bradbury et al. 2016;Cerezo-Román et al. 2017).
This paper investigates all available evidence for Middle Bronze Age cremations in Britain in order to evaluate the characteristics of cremation burial during the period and whether cremation cemeteries represented the majority of Middle Bronze Age people.It explores the size, duration, distribution, and associated structures of Middle Bronze Age cremation cemeteries, and subsequently analyses individual cremation burials, both in terms of their associated material culture and the osteological evidence relating to demography.Finally, this paper analyses the cremation burial process, encompassing the evidence for pyres, the burning of the deceased, and the collection and deposition of the remains.

URNS, INVADERS, & COMMUNITIES -PAST SCHOLARSHIP ON MIDDLE BRONZE AGE CREMATIONS IN BRITAIN
The earliest studies of prehistoric cremation burials in Britain focused on their chronology.Cremation burial in cemeteries was debated as being a Bronze Age tradition from the early 19th century (Miles 1826;Bateman 1861, 279-87;Lubbock 1865, 313;Stanley 1867;Pennington 1875).During this time, numerous sites containing urned and unurned, as well as single and multiple, cremation deposits were being identified (eg, Miles 1826; Bateman 1861).It was initially thought that urned cremation burials had been deposited in an earlier phase of the Bronze Age, although this was based solely on the relative paucity of their grave goods.These social evolutionary interpretative schema were soon supplanted by the development of a Bronze Age temporal framework based on pottery typo-chronologies.Abercromby's (1912) landmark corpus of British and Irish pottery formalised many Bronze Age pottery types and identified an 'overhanging rim' type, which later became the Collared Urn type (see Longworth 1961;1984) and the 'Deverel-Rimbury' forms (Abercromby 1912, 7-14).These were attributed to the Middle and Late/later Bronze Ages, respectively, and both contained cremation burials.The identification of these two ceramic types provided the foundation for all subsequent scholarship on Middle Bronze Age cremation cemeteries in Britain.As was typical in the late 19th/early 20th century, the cemeteries were interpreted within invasion-based paradigms (eg, Clay 1927;Kendrick & Hawkes 1932, 107;Childe 1947, 188).There is little osteological detail in the late 19th or early 20th century research on the cremated remains, many of which were not retained.
It was not until the advent of radiocarbon dating in the mid-20th century that the deposition of cremations within broader cemeteries was shown to pre-date the Urnfield funerary tradition in continental Europe (Ellison 1975, 373;Barrett 1976;cf. Sørensen & Rebay-Salisbury 2008).During the late 20th century, the greater chronological accuracy achieved through increasing radiocarbon dating and refining typochronologies (eg, M.A. Smith 1959, 155-9, 185;I.F. Smith 1961;Calkin 1962) did not substantially change the relative temporal sequence of funerary practices in Britain.Cremation cemeteries in Britain with Collared, and occasionally Food Vessel, urns were dated to the Early Bronze Age: Needham's Period 3 (2050-1700 cal BC) (Needham 1996, 130-2).Subsequently, the widespread use of flat and barrow sites with Deverel-Rimbury urned and unurned cremations occurred during the Middle Bronze Age: Needham's Period 4 and 5 (1700-1150 cal BC) (Needham 1996, 132-4).This Middle Bronze Age pottery and practice was suggested to continue until the Late Bronze Age (1150-800 cal BC) (Atkinson 1972, 115;Burgess 1980, 158-9;Brück 1995;Bradley 2007) at which time, as Brück (1995, 264) has demonstrated, cremations shift their location far closer to settlements and take on a different, less formalised role.

THE PREHISTORIC SOCIETY
It is notable that few works before the 1980s attempted a dedicated synthesis or explanation of Middle Bronze Age cremation burial practice in Britain.Instead, research identified the generalities of this burial form.In summary, burials were found with increasingly fewer grave goods, often in larger groups, and their pottery potentially represented certain groups with shared identities.Early contextual analyses were limited to suggestions of a preference for the placing of cremations in the southern section of barrow sites and the recognition that cremation cemeteries were found both in barrows and flat sites (Preston & Hawkes 1933).
It was only with the work of Burgess (1980) and Bradley (1981) that the landscape context of cremation burials was studied more thoroughlyalbeit with the analyses being published only in summary.These two scholars looked at the geographical location of Middle Bronze Age cremation cemeteries in southern England and concluded that many were found on good agricultural land and were geographically distinct to the burials of the Early Bronze Age Wessex tradition.This pattern was interpreted as a means to express property rights over preferable land 'at a time of more intensive farming' (Bradley 1981, 103).In contrast, Ellison (1980) focused on analysing the immediate character and locations of Middle Bronze Age cremations within the cemeteries themselves.This highly influential paper investigated the size, demographics, and funerary contexts of 48 multiple cremation sites, within a broader database of 480 sites and 608 urns across southern England, dated largely through the presence of Deverel-Rimbury pottery (Ellison 1980).However, the majority of the underlying data and data analysis was only available within her unpublished doctoral thesis (Ellison 1975) and, as with Bradley's (1981) paper, this relative inaccessibility has prevented a thorough re-evaluation of those original findings.Both Bradley's (1981) andEllison's (1980) research suggested that the rise of cremation cemeteries heralded a new social dynamic within Britain that lacked evidence of social distinctions.They highlighted the challenge of reconciling this funerary tradition with the settlement evidence and metalwork hoards that demonstrated new hierarchies (see papers in Barrett & Bradley 1980).
These two publications represent the last dedicated projects studying Middle Bronze Age cremations beyond site specific reports or county summaries.Subsequent scholarship appears in regional syntheses of Bronze Age burial practices (eg, Allen et al. 1987;Mullin 2003;Robinson 2007;Cooper 2016) or, in one case, as part of a broader synthetic analysis of later prehistoric funerary practices in southern Britain (Bristow 1998;2001).The publications emerging from the many Middle Bronze Age cremation cemeteries excavated since 1981 are, with the exception of the monograph on the site of Eye Kettleby, Leicestershire (Finn 2011), typically sections within broader monographs, site reports in county journals, or unpublished interim reports.Prior to this paper, no studies have re-analysed all sites across Britain excavated before and after 1981 together or challenged substantially the interpretations of Bradley (1981) and Ellison (1980).

METHODOLOGY
The purpose of this study is to reassess all Middle Bronze Age (c.1600-1150 cal BC) cremation sites in mainland Britain as well as those islands in close proximity, including the Isles of Scilly and Isle of Wight, the Western Isles, Shetland, and Orkney (Appx 1).The aim is to evaluate whether existing community-based models of cremation cemeteries are appropriate.
Initial data collection identified over 7000 cremation burialsdefined as a single cut feature containing cremated remains from one or more individualsfrom 1696 sites in Britain that had the potential to be Middle Bronze Age in date.This potential was based on the burial sites' features, material culture, or radiocarbon dates.Many sites had little or no diagnostic information to prove the sites' period of use.As such, a key challenge to this study was filtering this corpus down to only those sites where at least one cremation burial could be placed within the Middle Bronze Age with a high degree of confidence.
Of the 1696 cremation burial sites that had the potential to be Middle Bronze Age, 417 sites had at least one radiocarbon date associated with a cremation burial.Only 47 of these 417 sites contained burials that could be confidently assigned to the Middle Bronze Age through a direct radiocarbon date, rather than a phase that crossed into either the Early or Late Bronze Age.
It has been shown that radiocarbon dates obtained from cremated human bone can be offset by the inbuilt ages of the material used in the cremation pyre (Snoeck et al. 2014;2015).This effect can be negligible if the material used is of a similar age to the E. Caswell & B. Roberts. COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c. 1600-1150 CAL BC) individual, but can also distort the date of a cremation significantly in other cases, particularly if old wood, coal, or peat was used.Consequently, 21 cremation burials (at nine sites) were reclassified as Middle Bronze Age (their dates being only slightly older than 1600 cal BC), as there was also evidence for other Middle Bronze Age cremation burials on the site.This consideration in mind, the number of cremation burials sites containing at least one burial radiocarbon dated to the Middle Bronze Age rises to 56.
Of the 1696 potentially Middle Bronze Age sites, the remaining 1640 cremation burial sites which did not have at least one burial directly radiocarbon dated solely to the Middle Bronze Age were investigated for associations with typo-chronologically diagnostic material.There were 321 cremation burial sites that lacked sufficient diagnostic material and so were discounted from this study.The remaining 1319 contained at least one cremation burial which could be confidently dated to a typo-chronological range.There were 322 sites that contained at least one burial whose associated material is dated, according to current typo-chronological schemes, to the Middle Bronze Age.
When combined, the number of individual sites that contained at least one cremation burial confidently dated by radiocarbon (56 sites containing 673 individuals) or typo-chronology (322 sites containing 2460 individuals) to the Middle Bronze Age is 378 (see dating criteria details in Appx S1).These 378 sites contain 3133 individual burials, representing a minimum number of 3242 individuals that may be confidently dated to between c. 1600 and 1150 cal BC.These are the securely dated Middle Bronze Age cremation burial data which form the analytical core of this paper.Notably, these 378 sites also contained 1319 further burials which could not be confidently dated to the Middle Bronze Age.
The paper's results were created using a database designed to record each cremation burial site identified by their individual burials.This is accessible in the online supporting material (Appx S1).Where possible, each cremation burial was categorised according to: the funerary structure (eg, barrow, ring ditch, settlement, etc.); burial arrangement (eg, container, grave markings, grave goods etc.); and osteology (minimum number of individuals [MNI], sex, age, bone quantity, bone colouration), all of which were recorded using a standardised lexicon of terms.Each site was investigated as thoroughly as possible using the available records of each site.If these records did not provide information on the characteristics being recorded, the field was left blank rather than inferring a potential value.

Cemetery size
The 378 Middle Bronze Age cremation burial sites contain a mean number of 12 individual burials, dating to all periods, per site (Fig. 1).However, the mean is highly skewed by the presence of 19 (5%) large cemeteries containing over 50 cremation burials, such as at Simons Ground, Dorset (White 1982);Bromfield, Shropshire (Stanford et al. 1982;Hughes et al. 1995); and Vinces Farm, Essex (Erith & Longworth 1960) (Table 1).On the other hand, there are 210 (55%) sites that contain fewer than five individual burials.The number of burials recorded on these sites may be under-represented, as it has not been possible to confirm whether the excavated area of the cemeteries and the number of cremation burials excavated from them represent the full extent and number of burials placed on these sites.Having said that, there are still 117 (31%) sites that contain only a single cremation burial with only one individual, such as at Alwynds, Surrey (Germany 2010); East Harting Farm, Sussex (Aldsworth 1983); and Dishley Grange, Leicestershire (Walker 2009).This indicates that, despite the aforementioned biases against identifying and dating such sites, single burials remained a major feature of Middle Bronze Age funerary practices and that the majority of Middle Bronze Age cemeteries are far smaller than would be expected to contain an extended family.
Of the 3133 Middle Bronze Age cremation burials found, there are only 94 incidences (3%) of multiple individuals being placed within the same cremation burial (Table 2).The largest of these exceptional cremation burials with multiple individuals are: Shouldham, Norfolk, with a MNI of five (Wells 1976); Vinces Farm, Essex with a MNI of four (Erith & Longworth 1960); and Ring Ditch 4, between Linch Hill and Stanton Harcourt, with a MNI of three (Hamlin & Case 1963).
The utility of the MNI when studying cremation burials is partially restricted due to the fragmented nature of the remains, which may lead to the identified number of individuals being lower than the actual number deposited in the past.Similarly, it is equally possible that a single cremated individual may have had THE PREHISTORIC SOCIETY their remains separated into more than one token deposit (see commentary in Lynch & O'Donnell 2007, 110).These caveats aside, it is proposed that, regardless of the size of the cemetery, Middle Bronze Age cremation burial practice overwhelmingly favoured the burial of single individuals within burials.This is significant, as the presence of burials with multiple individuals shows that cremation easily facilitates the mixing of human remains within relatively smaller receptacles than might otherwise be required.Similarly, one can envisage the possible symbolic gestures that might be achieved by the mixing of a homogenised cremation material.Yet, these mixing options were almost always rejected in the final deposition of human remains.

Cemetery duration
There are 97 Middle Bronze Age cremation burial sites which have at least one radiocarbon date, although only 56 of these sites' radiocarbon dates place individual cremation burials within the Middle Bronze Age.Unfortunately, too few sites have suitable spatial and temporal details to allow any comment on the shifting use of cremation cemeteries as seen elsewhere in Europe (eg, De Reu et al. 2012).
Only 30 of the 97 cremation burial sites have at least three separate burials that have been radiocarbon dated.Three of these sites have reports that use Bayesian modelling (citing Buck et al. 1996) to determine their duration.At Papworth Everand Bypass, Cambridgeshire (Hounsell 2007, 20), this modelling determined that the cemetery's span of use was between 1 and 140 years.Assuming that the radiocarbon dates reflect the full span of use of the cemetery and that the cremation burials were evenly placed through time, this would require at least one cremation burial to be placed within the cemetery every 2 years (minimum number of burials: 57; MNI: 67).At Eye Kettleby, Leicestershire (Finn 2011, 56-8), the duration of all funerary activity is modelled to be between 220 and 400 years.Following the same assumptions as stated above, this would require at least one cremation burial to be placed within the The size of the 378 cremation burials sites when considering all cremation burials found on the site compared to the size of the 378 cremation burials sites when considering only those cremation burials that could be confidently dated to the Middle Bronze Age The remaining 27 Middle Bronze Age cremation burial sites with three or more radiocarbon dates (minimum number of burials: 1027; MNI: 1073) lack any modelling, and therefore comment on the length of use of these cemetery sites can only be generalised.The average maximum span of use of these 27 sites is 866 years, such that on average one burial must have been placed across these sites every 59 years.This contrasts strongly with these same sites' minimum average span, which is only 200 years, such that on average one burial could have been placed in these sites as regularly as every 12 years.Without modelling, the maximum length of these sites' duration, seen purely from the radiocarbon dates, is likely to be exaggerated, as will the number of years between the placement of each burial at these locations.Equally, it is unlikely the 12 sites (or 14 if including those sites that have used Bayesian modelling) which have radiocarbon dates (Table 3) that would allow for their cremation burials to be buried within a single year were used for so short a time.It does appear that the larger cemeteries' span of use is extended by the presence of a minority of burials dating either earlier than the majority of the burials, such as at Eweford West, East Lothian (MacGregor 2008) and Moverons Pit, Essex (Clarke & Lavender 2008), or later than the majority of the burials, such as at Western International Market, London (Bradley 2003;Boyer 2007); Handley Hill, Dorset (Barrett et al. 1981); and Simons Ground, Dorset (White 1982).
Yet, there are also, generally smaller, cemeteries which show a low number of burials that were placed intermittently over a long period of time, such as at Whitton Hill Henge, Northumberland (Fowler 2013); Heathrow Terminal 5, London (Framework Archaeology 2010); and Dallam School, Cumbria (Platell et al. 2013).Similarly, of these 27 sites with three or more radiocarbon dates, there are at least seven cremation burial sites that must have been used over more than 200 years.As such, it is clear that there is no set rule for the length of cemetery use or the regularity at which burials were placed at these sites.

Distribution
The Middle Bronze Age cremation burial sites included in this study are distributed throughout mainland Britain and the nearby islands (Fig. 2).They are found far more frequently in southern England, and in particular are most densely concentrated around the south Dorset and Hampshire region, but are strikingly and inexplicably absent from the High and Low Weald of East Sussex, an absence that is paralleled in the Bronze Age settlement record (Caswell 2018).There are relatively few sites in Wales, although this might be due to the relatively lower level of fieldwork and lack of sufficiently well dated sites.There are notably few cremation burial sites in northern Britain, particularly in the region spanning north-east England and southeast Scotland (Warden et al. 2016).When the distribution is analysed from the perspective of cemetery size, the most northerly Middle Bronze Age cremation burial cemetery with over 50 individuals is at Bromfield, Shropshire (Stanford et al. 1982;Hughes et al. 1995).Whilst the uneven distribution of excavations and research needs to be considered (see Green et al. 2017), this north-south division is nonetheless striking.

Funerary structure
Middle Bronze Age cremation burials are found primarily in three funerary contexts: barrows (195 sites, 52%), ring ditches (60 sites, 16%), and flat sites (defined as having no features visible on the ground surface) (122 sites, 32%).It is certainly possible that a proportion of the ring ditches are the remains of barrows which have been subsequently ploughed out.Sixteen (<4%) sites contained detail in their reports suggesting an association with field/enclosure features, and 36 (<10%) made mention of nearby settlement features, although these were placed up to 1 km away.
The 3133 cremation burials within the 378 cremation burial sites reveal very limited evidence for the marking of burials with visible, above ground, markers.For example, there is evidence at a few sites such as Itford Hill, Sussex (Holden 1972), where each cremation burial had an associated post-hole which was interpreted as a marker to prevent intercutting.However, the evidence for the presence or absence of grave marking is not consistently interpreted and can be difficult to identify (see Evans & Knight 1998).
There are examples where both intercutting and the lack of intercutting have been interpreted as indications of above ground grave markers.At Briar Hill, Northamptonshire (Bamford 1985), the cremation burials were found to intercut with one another, leading to an interpretation that their locations must have been marked.At Broom, Bedfordshire (Cooper & Edmonds 2007), four cremation burials were each associated with a post-hole, yet were intercut by other cremation burials, leading to an interpretation for the absence of grave marking.

Funerary containers and grave goods
Of the 3133 Middle Bronze Age cremation burials, 844 (27%) show no evidence for a burial containerwhether ceramic vessel, stone cist, or wooden coffin and a further 211 (7%) have no record of this information in their reports (Table 4).Of the remaining 2078 cremation burials, 2058 (66%) were excavated in a container, 2036 of which are ceramic vessels.Where the information for position of the ceramic vessel holding the individual cremation burials was available, it showed that ceramic vessels were almost evenly found inverted (460 burials) and upright (485 burials).Thirty-five sites had a mix of individual cremation burials in both inverted and upright ceramic vessels, indicating that the choice of urn direction was not dictated by burial site.Finally, there are 20 burials where the clustering of the cremated bone and ash led to the inference of an organic bag, such as at Latch Farm, Kent (Piggott 1938); Briar Hill, Northamptonshire (Bamford 1985); and Eye Quarry, Cambridgeshire (Patten 2004;2009).
There are 1003 cremation burials (32%) which contained no surviving accompanying artefacts (such as ceramic vessels, flint, metalwork, or animal bone) that might be regarded as grave goods, although animal bone may be under-represented due to the difficulties Of these 137 burials, there are 43 cremation burials from 23 cremation burial sites which contained burnt objects, presumably also gathered from the funerary pyre.These were predominately flints, such as at Itford Hill, Sussex (Holden 1972); Game Farm, Suffolk (Gibson 2004); Barnes Urnfield, Isle of Wight (Dunning 1931); and Pokesdown, Kent (Clay 1927).There are also examples of: molten bronze as at Eye Kettleby, Leicestershire (Finn 2011); animal bone as at King's Hill, Bedfordshire (Cooper & Edmonds 2007); and flint arrowheads as at Standlake ring ditch 1, Oxfordshire (Riley 1946) and Colne Fen, Cambridgeshire (Evans & Appleby 2013).In 24 of these 43 cremation burials, the burnt objects were found associated with ceramic vessels, which suggests that these objects were selected for inclusion in the burial.

Demographics
In the last three decades, osteological analyses have typically followed the framework outlined by McKinley (1997;2000).However, before the 1990s, osteological reports were far more varied, such that it is not uncommon for certain traits of skeletons or whole assemblages to be unrecorded.Osteological analyses have allowed estimation of the sex of 196 individuals from 190 cremation burials across 70 sites, and the age ranges of 991 individuals from 884 cremation burials across 150 sites.In total, 840 (85%) of the burials that were aged and 159 (81%) of the burials that were sexed were detailed in reports published in or after 1980 (Figs 4 & 5) and are therefore regarded as reliable for the purposes of this study.
Of the 154 individuals where sex has been estimated, only 54 individuals have been confidently sexed as females (30 individuals) and males (24 individuals), an almost even division.In the remaining 104 individuals where the sex estimation is less certain, there is a slightly higher number of possible females.
Of the 740 individuals that have an osteologically determined age range, all osteological age ranges are represented, but 414 (49%) of the individuals are adults (25-40 years).It is possible that children (3-11 years) and infants (0-2 years) are under-represented due to their smaller bone size or preferential destruction rate for this age group.Recent experimental research suggests cremated child remains should remain archaeologically visible (Holck 1997;Jaeger & Johansen 2013); therefore, the low proportion of burials might reflect a cultural bias (Lewis 2007).Whilst these age ranges and categories can be osteologically determined, it is acknowledged that the social recognition and relevance of these ages could well have been different in the past.These caveats mean that any statistically significant correlation between age range or sex and the presence or absence of grave goods (beyond ceramic vessels) found in Middle Bronze Age (c.1600-1150 cal BC) cremation burials in Britain should be treated with caution.
However, it is worth stating that individual infant cremation burials are treated in the same or similar manner to adult cremation burials found elsewhere, such as at Butcher's Rise, Cambridgeshire (Evans & Knight 1998); Aldham Mill Hill, Suffolk (Everett & Boulter 2010); and King's Dyke, Cambridgeshire (Knight 1999).Whilst there are probable female individuals buried with infants, such as at Broad Chalke 1, Wiltshire (Grinsell 1957) and Oliver's Battery, Hampshire (King 1989), there are also probable male individuals buried with an infant, such as Burial 2257 at Papworth Everard Bypass, Cambridgeshire (Hounsell 2007;Gilmour et al. 2010).As such, it is clear that neither age nor sex provided an insurmountable barrier to the various forms of cremation burial practiced during the Middle Bronze Age.
Burning the bodies: Pyre evidence The colour of cremated human bone was recorded for 306 cremation burials with the vast majority being described as 'well calcined' or 'buff white'.This indicates that human bone was burnt in pyres with temperatures exceeding 600°C (Shipman et al. 1984).The complexities involved in the preservation and identification of Bronze Age cremation pyres in Britain have recently been addressed systematically and in depth in three experimental replications inspired by the excavation of pyre evidence under the Early Bronze Age barrow at Guiting Power 3, Gloucestershire (Marshall 2011).Pyres are typically inferred from evidence such as burial scorching, where the soil surrounding the cremation burial has been burnt, such as at Claggan, Arygll (Ritchie & Thornber 1977); Zionshill Copse, Hampshire (Entwistle 2001);and Kalis Corner, Hampshire (Dacre & Ellison 1981).It has also been inferred from the burning of pottery, suggesting that the ashes were deposited when still hot, such as at Imperial College Sports Ground, London (Powell et al. 2015); Temple Guiting 8, Gloucestershire (O'Neil 1967); and Swanmore, Hampshire (Dunning 1931).Of the 378 Middle Bronze Age cremation burial sites, there are 43 that have in situ evidence for burning that indicates either that a pyre existed at the site, or that the remains of a pyre were deposited at the site whilst still in a state of combustion, implying that the original pyre had been nearby.
Considering that 3133 cremation burials have been studied, it is reasonable to expect that more evidence for pyres would have been identified, assuming that cremated remains were not being brought to their place of burial from elsewhere.Furthermore, it would be expected that if in situ evidence for burning had existed at these sites, it would have been found in larger quantities, especially at the largest cremation burial sites.It has been suggested at sites such as Eye Kettleby, Leicestershire (Finn 2011); Butcher's Rise, Cambridgeshire (Evans & Knight 1998); and Mockbeggar Lane, Hampshire (Coles 2004), that funerary pyres were potentially cleaned away after their use due to the high collection rates of bone excavated in the cremation burials.However, there are 32 pyre sites that exist at cremation burial sites; these are invariably simple spreads of ash and charcoal, which show no signs of cleaning beyond the collection of human remains.

COLLECTING AND DEPOSITING THE CREMATED DEAD
Whilst taphonomic issues should not be ignored, the published data reveal that there was no standard practice for the quantity of human remains that underlie their collection and burial during the Middle Bronze Age in Britain, as there are significant variations from every perspective.
There are 859 cremation burials from 104 cremation burial sites which recorded the cremation deposit weight of the excavated human bone (Fig. 6).The mean cremation deposit weight of human bone from these is 374.6 g, although it should be noted that truncation might have reduced the quantity of bone recovered at certain sites.These weights vary significantly (with a standard deviation of 561.7 g) (Table 5).For instance, 635 (74%) of the cremation burials weigh less than 500 g, and 370 (43%) cremation burials weigh less than 100 g (Fig. 7).Only 17 (2%) cremation burials weigh over 2 kg.The most widely cited experimental study which recreated Anglo-Saxon pyres found that the cremation of an adult could be expected to produce between 1.5 and 2 kg of bone residue (McKinley 1997).
The extensive presence of cremated human remains weighing significantly less than 1.5-2 kg implies the partial collection of human remains for burial.Unfortunately, too few osteological reports recorded the cremation deposit weights by fragment size, as recommended by McKinley (1997;2000) to enable further comparative analysis.An alternative method based on the volume of cremated human remains has been proposed that seeks to provide a more reliable representation of the completeness of the human remains that were originally buried (Harvig & Lynnerup 2013).However, none of the publications on Middle Bronze Age cremations in Britain used by this paper employed this new methodology.
When the cremation deposit weights within 66 cremation burial sites are investigated, 56 sites show a relatively higher internal consistency than the total population of weighed burials.For instance, within the 32 cremation burials at Kingsborough, Kent (Allen et al. 2008) and the 11 cremation burials at Coton Park, Warwickshire (Maull 2001), there is a remarkably consistent selection of less than 100 g in the quantity of bone being buried, which cannot easily be explained by issues of taphonomy and differential recovery.Such consistency has also been recognised within separate clusters of potentially chronologically distinct cremation burials (Finn 2011, 66).Other Middle Bronze Age funerary rites in Britain are evidenced, although these comprise less than 3% of the total number of burials from this period.A new comprehensive survey of non-burnt human remains dating to the Middle Bronze Age has only identified 92 individual burials from 45 sites which emphasises the dominance of cremation based practices at this time (Cormack 2018).The sites include inhumations in bogs, as at Ashton Moss, Lancashire (Mullin 2003;Nevell 2015); in rivers, as shown by the dating of skulls (Bradley & Gordon 1988;Schulting & Bradley 2013); in water holes, as at Striplands Farm, Cambridgeshire (Evans & Patten 2011); in ditches, as at Tormarton (Osgood 2006); and settlements, as at Gwithian, Cornwall (Nowakowski 2004, section 3, 26, appx 11).Many sites with inhumations also contained cremated human bone, which is sometimes in greater abundance than the unburnt bone, such as at Weymouth 34, Dorset (otherwise known as the famous Rimbury cemetery that defines the Deverel-Rimbury ceramic type) (Warne 1866); Milborne 16h/i, Dorset (Grinsell 1959); and Berwick St John 10, Wiltshire (Grinsell 1957).Within our study, cremated and unburnt human bone was found at 40 of the 378 THE PREHISTORIC SOCIETY cremation sites.This indicates that these cemeteries were not exclusively used for one funerary rite, although cremations form the vast majority of the Middle Bronze Age funerary remains visible in the archaeological record.In summary, the archaeologically visible cremation burial practice in Middle Bronze Age Britain can be characterised by: the placing of individuals of both sexes and all ages; variable quantities of bone; associated with few grave goods, barring ceramic vessels; either in unmarked flat sites or earlier monuments.

Is there a Middle Bronze Age cremation horizon in Britain?
There is no easily identifiable temporally distinct appearance of a Middle Bronze cremation tradition at 1600 cal BC.Brück (2014, 130) asserts that the apparent dominance of cremation burials over inhumation burials occurs in Britain from c. 2000 cal BC.
A similarly placed transition in north-east Englandsouth-east Scotland is also suggested by Fowler and Wilkin (2016, 126) at 1900 cal BC.Needham's chronology (2011; see also Needham et al. 1997) for the entirety of Britain recognises that a transition is hard to pinpoint, suggesting that the cremation rite becomes more frequent towards the end of his phase two (2300-1950 cal BC) and only becomes the 'predominant burial rite' by his phase three (1950-1500 cal BC).
There are numerous radiocarbon dated funerary sites, containing multiple cremation burials, whose use spans the 2nd millennium BC.These include the sites of: Meldon Bridge, Peebleshire (Speak & Burgess 1999) Needham et al. 2013).The main explanation for this lack is that the characteristics of the surviving funerary evidence from 1600 cal BC in North-west Europe are less conducive to broader comparative analyses.In addition, the complexities and different sub-divisions of typochronologies and their respective terminologies can prove obstructive (see Roberts et al. 2013), as can the varying practices of archaeological fieldwork in the region (eg, Webley et al. 2012).However, the similarities and differences in mid-late 2nd millennium BC funerary practices have been discussed in the broader context of North-west European later prehistoric archaeology (eg, Bourgeois & Talon 2009, 39-42;Bradley et al. 2015, 195-205;Marcigny et al. 2015, 231;Webley 2015).
The closest and most widespread parallels to the Middle Bronze Age cremation burials in Britain are found to the west, in Ireland.The wealth of recent excavations and radiocarbon dates means that the E. Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c. 1600-1150CAL BC)  funerary framework proposed by Grogan (2004), of a phase (c.1500-1300 cal BC) of burials placed with Cordoned Urns and grave goods followed by a phase (c.1300-1000 cal BC) of cremations placed in coarser urns, is in need of substantial chronological revision.The recent modelling of radiocarbon dates by Brindley (2007) and Bayliss and O'Sullivan (2013) has confirmed that the use of Cordoned Urns had ceased by c. 1500 cal BC.It also appears that razor knives, faience, gold, and amber grave goods were only included in burials pre-dating 1600-1500 cal BC (Waddell 2010).
A recent review of all accessible reports on later Bronze Age (c.1600-600 cal BC) cremation burials found along road schemes in Ireland has found that following c.1600-1500 cal BC, the evidence of funerary practices is dominated by small cremation deposits, usually unaccompanied, but sometimes placed within coarse ceramic vessels (Spillane 2017).Other than ceramic vessels, or later typically only ceramic sherds, the objects placed with these cremations included only the occasional burnt or unburnt flint flakes and animal bones.These burials were predominantly found in flat, unmarked pits, either as isolated features or in small clusters (Lynch & O'Donnell 2007;McQuade et al. 2009, 141-6;Troy 2015;Cooney 2017).
These Irish cremation burials are placed in similar locations to those seen in Britain, having been found deposited in barrows, ring-ditches, and occasionally in close proximity to Bronze Age settlements (Spillane 2017).In some cases, it seems as though cremation depositions in the Middle Bronze Age became focal points for domestic settlement in the Late Bronze Age (Spillane 2017, 47-8).Furthermore, and as seen in Britain, the majority of Middle Bronze Age cremation burial sites contain only a few individuals, though there are occasional larger cemeteries, such as at Templenoe and Derrybane, Co. Tipperary; Manusmore, Co. Clare (Bermingham et al. 2012);and Rathglass, Co. Galway (Doody 2008;Kiely & O'Mahony 2011;Péterváry 2009).These larger cemeteries are generally specific to the Middle Bronze Age, after which isolated pit burials become the dominant form (Spillane 2017, 37-8).
Middle Bronze Age pyres are rarely associated with these burials, although some examples such as Newford, Co. Galway or Coolmore, Co. Kilkenny contain pit features with indications of intense in situ burning, along with small deposits of burnt bone, which could be an indicator of pyre structures.Such evidence has led Becker (2014, 14) to suggest that the small amounts of cremated remains located near these pyres represent the intentional deposition of pyre material rather than formalised burials.
In northern France and southern Belgium, cremation cemeteries containing multiple individuals, some contained within ceramic urns and a few accompanied by additional grave goods, are found from c. 1500-1100 cal BC (Le Goff & Guichard 2005;Bourgeois & Talon 2009;Le Goff & Billand 2012).The current dating of evidence across northern France, Belgium, and the Netherlands indicates that few barrows were constructed from c. 15/1400-1100 cal BC, with cremations from this time being inserted into older barrows, such as at Waben-Le Sémaphore (Desfossés & Bernard 2000; see also Bourgeois & Arnoldussen 2006;Bourgeois & Fontijn 2008;Bourgeois & Talon 2009;Bourgeois 2013).Given the similarities in the ceramic assemblages on either side of the Channel (see Marcigny et al. 2007;Kleijne 2010), it is not surprising that a ceramic urn, closely related to the Deverel-Rimbury ceramic types, containing a cremation was found at Argoeuves-Le Moulin d'Argoeuves (Soupart 2009).
Further inland into Continental Europe, a section of the Seine valley has one of the most extensively excavated mid-late 2nd millennium BC funerary landscapes in North-west Europe.It contains a diversity of barrow monuments, co-existing inhumation and cremation funerary traditions, and a wide range of grave goods, providing a cautionary example to any straightforward narrative attempted in less well excavated areas (Delattre & Peake 2012;Delattre et al. 2015;Rottier 2010).To the west, Middle Bronze Age funerary sites in north-west France are also dominated by barrows containing (where their preservation in the soil allows) both cremations and inhumations with ceramic urns and few grave goods; their chronologies remain poorly understood (Briard 1984;Fily et al. 2012;Boulud-Gazo et al. 2017).
Funerary sites in north Germany and Denmark include occasional cremation cemeteries, such as at Lustrupholm, Denmark (Feveile & Bennike 2002), on flat sites that display similarly equal proportions of ages and sexes to Middle Bronze Age cremation cemeteries in Britain.However, the vast majority of the archaeologically visible contemporary funerary activity comprises thousands of barrows, inhumations, and grave goods (see Bergerbrant 2007;Holst & Rasmussen 2013).

THE PREHISTORIC SOCIETY
The existence of shared practices connected by maritime routes throughout North-west Europe during c. 1600-1150 cal BC is not in doubt, as a major museum exhibition, BOAT 1550 BC (Lehoërff 2012), and numerous recent conferences have demonstrated (see Bourgeois & Talon 2005;Clark 2009;Lehoërff & Talon 2017).However, the extent of similarities and differences in funerary practices has remained underinvestigated and deserves further attention beyond the scope of this paper.

Identifying long term trends in the quantity of archaeologically visible burials throughout the Bronze Age in Britain
There is a substantial decrease across Britain in the number of funerary sites in the Middle Bronze Age (c.1600-1150 cal BC) as compared to the Early Bronze Age (c.2200-1600 cal BC), which is not adequately explained by its only slightly shorter duration (150 years difference).For instance, in the Tyne-Forth region spanning north-east England and south-east Scotland, there are over 130 Early Bronze Age sites (Fowler & Wilkin 2016), three Middle Bronze Age cremation funerary sites, and 12 Late Bronze Age burial sites, only six of which contained confirmed cremation burials (Warden et al. 2016).Similarly, in Cumbria, north-west England, 100+ funerary sites have been dated to the Early Bronze Age, yet only one Middle Bronze Age funerary site has been confirmed (Evans 2008;Platell et al. 2013, 53;Walsh 2013).Even in areas where there are relatively high concentrations of Middle Bronze Age cremation burials, such as in the Dorset-Hampshire region, there are far more Early Bronze Age burial sites (Grinsell 1959;Bristow 1998;2001).
The Late Bronze Age (c.1150-800 cal BC) funerary record (inhumations and cremations) in Britain is even sparser than that of the Middle Bronze Age.In southern Britain, Late Bronze Age cremation burials are predominately found in roundhouses and associated ditches (see Brück 1995;Roth 2012;Davies 2016), comprise smaller quantities of human bone, and are rarely accompanied by ceramic vessels.In northern Britain, similarly small quantities of cremated and unburnt human bone are found in settlements, caves, ditches, and in both earlier and contemporary funerary monuments (Thomson 2011;Melton et al. 2016;Warden et al. 2016).The cremated remains in these studies (referenced above) also appear to be generally smaller in cremation deposit weight than the average Middle Bronze Age cremation burial.This reduction in the quantity of cremated bone being buried through time has also been demonstrated across the Early, Middle, and Late Bronze Age in Cambridgeshire (Evans 2016, 429).

Reassessing the Middle Bronze Age community cemetery model in Britain
When considering all Middle Bronze Age cremation burial sites with at least three radiocarbon dates, it is possible that cremations need only have been deposited on each site on average once every 54 years (Table 3).The quantity of cremation burials can also be analysed on a national scale.The period of study encompasses 450 years such that, from the available evidence, only one archaeologically known cemetery site was created every 2 years.Furthermore, only seven deaths a year would then have resulted in a cremation burial being placed in one of these sites.While these numbers are likely to under-represent the number of individuals who were cremated and buried, it can only be concluded that a small minority of people in Middle Bronze Age Britain were buried in an archaeologically visible rite.For instance, the placing of cremated human remains in seas, rivers, lakes, and bogs or their scattering over the landscape would all be invisible to later archaeological investigation.
When compared to the contemporary settlement record, how few individuals received an archaeologically visible cremation burial is even clearer.There are c.8000 substantial domestic sites in Britain, characterised by the presence of at least one roundhouse, which are dated to the Middle Bronze Age (1600-1150 cal BC), either by radiocarbon dating, associated material culture, or architecture (Caswell 2018).This contrasts with the 3133 cremation burials dated using the same methods identified in this paper.Even when allowing for the fragmentary evidence of other funerary rites, the comparative ratio is stark; there are 2.6 settlements for each cremation burial in the Middle Bronze Age in Britain.Even considering the multiple taphonomic issues, and the many more sites that are doubtless awaiting re-dating or discovery, these statistics must lead to a revision of the community cemetery model applied to all Middle Bronze Age cremation burials in Britain.
Middle Bronze Age funerary sites containing cremation burials are invariably interpreted as cemeteries for E. Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c. 1600-1150CAL BC)  nearby communities (eg, Ellison 1980;Bradley 1981;Boyer 2007;Cooper & Edmonds 2007;Finn 2011) due to two influential publications (Ellison 1980;Bradley 1981) which relied upon two observations.Firstly, the size and clustering of these sites is similar to the assumed extended family unit of the Middle Bronze Age (Ellison 1980).The proposed spatially distinct clustering of cremation burials in groups of between ten and 30 individuals, as identified by Ellison (1980, 122), suitably fits the estimated size of extended family units that are widely thought to be the social core of Middle Bronze Age societies.This identification of clusters has been made in many subsequent site monographs such as Eye Kettleby, Leicestershire (Finn 2011); Daneshill, Hampshire (Millett & Schadla-Hall 1992, 91); and Oliver's Battery, Hampshire (King 1989, 22).However, this clustering is often only loosely defined methodologically and interpretatively, such as at Pasture Lodge Farm, Lincolnshire (Allen et al. 1987, 210), and when not conforming to the size suggested by Ellison, the clustering is instead defined by smaller groupings of three per cluster, such as at Papworth Everand, Cambridgeshire (Gilmour et al. 2010, 22).A detailed analysis of 60 Middle Bronze Age cemeteries in East Anglia by Robinson (2007, 51) found no evidence for clusters of cremation burials.
This study has shown that the vast majority of Middle Bronze Age cremation burials (2942, 94%) are found on a site with at least one other burial also dated to the period.However, it has been possible to confirm that a slightly larger proportion of Middle Bronze Age cremation burial sites (79 out of 118 cremation burial sites with only one burial, 21%) definitely contain only one cremation burial, while only 40 out of 378 (11%) Middle Bronze Age cremation burial sites contain 20+ cremation burials.This is despite the biases against recovering and dating single cremation burials, which makes it likely that the overall proportion of single cremation burials is under-represented.As such, despite Middle Bronze Age cremation burials frequently being found together, the majority of cremation cemeteries do not appear to meet the underlying requirement for demonstrating evidence for extended family units and therefore a larger community.
Secondly, Middle Bronze Age cremation burial sites are placed near, and have been linked to, contemporary settlements (Bradley 1981).The pairing of Middle Bronze Age settlements and cemeteries in Britain is frequently asserted (eg, Bradley 2007, 185;Darvill 2010, 222), usually on the basis of spatial proximity, such as at Down Farm, Cranborne Chase (Barrett & Bradley 1980).There is also the frequent comparative analogy to Itford Hill, Sussex (Ellison in Holden 1972, 110) where two sherds of pottery were found to re-fitone from a Middle Bronze Age barrow and the other in a nearby Middle Bronze Age settlement.The spatial proximity argument for pairing settlements with cremation cemeteries is based primarily on Bradley (1981), who asserts that the majority of Middle Bronze Age cremation cemeteries are found within 700 m of a settlement, with a peak between 50 m and 300 m (Bradley 1981, 100).
A pilot assessment of this assertion was made possible by comparing 372 of the 378 Middle Bronze Age cremation burial sites to the location of potential Bronze Age settlements known to all Historic Environment Records within mainland England, Scotland, and Wales, a database totalling 21,831 sites.The list of settlements (Caswell 2018) includes sites that might be Early or Late Bronze Age and some sites which are only speculated as being Bronze Age.It is therefore a generous distribution that would be expected to skew results towards a smaller distance between Middle Bronze Age cremation burials and Middle Bronze Age settlements.Six of the 378 Middle Bronze Age cremation burial sites were excluded from this analysis due to the poor settlement evidence available for their region.
This analysis found that there is a peak of 96 (26%) Middle Bronze Age cremation burial sites placed within 300 m of a potential Bronze Age settlement sitein both northern and southern Britainwhich might in part support Bradley's assertion that, in some cases, settlements are paired with cemeteries (Fig. 8).However, only 139 (37%) cremation cemetery sites were found within 700 m of a potential Bronze Age settlement, 201 (54%) cremation cemetery sites were located over 1 km away, and the average distance between these cemeteries and their nearest potential settlement was 1787 m.Therefore, from the data available to this paper it can only be concluded that Middle Bronze Age cremation burials do not show a universally strong spatial connection to occupation sites as has been suggested (Darvill 1996, 116-17;Bradley 1981;2007, 185).
Furthermore, the contemporary chronology of settlements and nearby cemeteries is often assumed rather than demonstrated.Yet, similarly to cremation sites in Ireland (Spillane 2017), when the radiocarbon THE PREHISTORIC SOCIETY dates for well excavated Bronze Age settlements and cemeteries within 500 m in Britain are compared, they frequently reveal that it is Late Bronze Age settlements that are placed in close proximity to pre-existing Middle Bronze Age cremation cemeteries.This occurs at Dunch Hill, Wiltshire (Andrews 2006); Game Farm, Suffolk (Gibson 2004); and Biddenham Loop, Bedfordshire (Luke 2008) and has also been observed across the Netherlands (cf.Gerritsen 2007).Directly contemporary Middle Bronze Age settlements and cemeteries in close proximity, such as at Shorncote Quarry, Gloucestershire (Barclay et al. 1995), are very rare according to the radiocarbon dates, are rarely discussed in site reports, and, bar the much-cited Itford Hill example (Holden 1972), never directly evidenced through material culture.
As such, it is argued that neither Ellison's (1980) nor Bradley's (1981) observations can now be broadly supported.Consequently, the community cemetery model in Middle Bronze Age Britain should be revised.
How then should Middle Bronze Age cremation burials be understood?The preference for burying only individuals does suggest that the deceased individual was given some primacy during multiple stages of the funerary process.However, this primacy did not translate to the final burial stage, which tended to avoid both monumentalisation in the landscape or any material expression of identity beyond a ceramic vessel that frequently contained their remains.This relatively homogeneous and modest burial stage may well have extended to the burning of the individual, given the absence of burnt personal ornaments which are known to have been worn but are deposited unburnt elsewhere in Britain (Roberts 2007).On average, less than three-quarters of the human bone produced through the cremation process is recovered from Middle Bronze Age cremation burials.Given the low number of sites showing evidence for cremation pyres, it seems likely then that the human cremated bone was 'created' in a different location to the final burial place and that the majority of it was used and/or placed in ways that are no longer archaeologically visible.The intentions underlying these uses will be difficult to understand, as they are likely to have occurred in the transportation (and maybe redistribution) of the potentially symbolically charged cremated material in social contexts which now leave no archaeological trace (cf.Appleby 2013;Kuijt et al. 2014;Bradbury et al. 2016).This understanding does not exclude Ellison's (1980) suggestion that cemeteries or grouping of burials represent kin groups, yet it should be stressed that the evidence above implies that the cre-  Similarly the regular use of Bayesian modelling and more substantial dating programmes on all cremation burial sites will allow a truer image of their life histories to be produced.There are substantial issues with the typo-chronologies of (Early-) Middle Bronze Age ceramics associated with cremations which would benefit from further study and dating.The recorded diversity of funerary practices indicates that there may well be far more sites than is currently appreciatedespecially when funerary sites and burials assumed to be Early Bronze are taken into consideration.However, even if all of these burials were dated to the Middle Bronze Age they would still be far exceeded by the number of contemporary settlements and would thus still represent only a minority of the contemporary population.This paper challenged the widely held assumption that Middle Bronze Age cremation burials represented entire communities (Ellison 1980) who were locally based (Bradley 1981).In this, it follows an earlier comparative analysis of Bronze Age settlements and funerary sites in the Netherlands, which also challenged successfully similar pre-existing models (Bourgeois & Fontijn 2008).This raises a fundamental questionwhere did the people who were subject to cremation burials in Middle Bronze Age originate?
The recent successful application of strontium isotope analysis to cremated burials (Snoeck et al. 2015) now enables this question to be at least partially resolved by demonstrating whether those individuals buried had lived and died nearby and whether there is a coherency in the life histories of the dead within a cemetery.The recent results from the isotopic analysis of preceding Beaker burials spanning the Chalcolithic and Early Bronze Ages (c.2500-1600 cal BC) highlight the co-existence in funerary treatment and contexts of both sedentary and mobile individuals, albeit within Britain (Parker Pearson et al. 2016).However, this has recently been challenged by the analysis of Neolithic-Early Bronze Age human aDNA which strongly implies a continental migration of people from Continental Northwest Europe to Britain, apparently replacing nearly the entire indigenous population, during the same period (Olalde et al. 2018).The widespread adoption of cremation-orientated burial practices prevents any extension of comparably extensive aDNA research programmes into the Middle-Late Bronze Age (c.1600-800 cal BC) in Britain.The materials and technologies in the Middle Bronze Age (c.1600-1150 cal BC), most visibly in bronze and gold as exemplified by the Salcombe, Devon and Langdon Bay, Kent shipwrecks (Needham et al. 2013;Wang et al. 2016), would imply continuity in mobility across North-west Europe and beyond.However, the current interpretations surrounding the contemporary construction of roundhouses, enclosures, and field systems in Britain, both implicitly and explicitly, strongly envisage far more sedentary farming groups.The resolution of this major interpretative divergence on the same prehistoric population (cf.Roberts 2013), ideally through the widespread application of strontium isotope analysis on cremated human bones, would enable a far clearer understanding of the extent to which mobility shaped and defined Bronze Age communities in Britain.
Strasbourg, France (June 2014).We are very grateful for the invitations to present and for the subsequent discussions with conference participants.We would like to thank Anna Bloxham, Emily Carroll, Claudio Cavazzuti, Kerri Cleary, Anwen Cooper, Katie Cormack, Chris Fowler, Arjan Louwen, Mette Løvschal, Rebecca Peake, Ben Spillaine, and especially Ann Woodward (née Ellison) for reading and commenting on an earlier draft of this paper.We are especially grateful to the three reviewers and the editors.Any errors remaining are the responsibility of the authors.

SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit: https://doi.org/10.1017/ppr.2018.9sourced.The details included within these reports were then recorded in a relational database (Microsoft Access 2013).Any further sites cited in these reports that had the potential to contain a Middle Bronze Age cremation burial were also added to the database and recorded when relevant.

Data selection criteria
Radiocarbon: Those sites with radiocarbon dates were evaluated according to: whether their absolute calibrated date range at 2σ lay solely between 1600 and 1150 cal BC; or where dates provided a terminus ante or post quem that overlapped 1600 cal BC or 1150 cal BC.Those sites with burials meeting these criteria were assigned to the Middle Bronze Age, while those sites with burials whose radiocarbon dates only overlapped with either the 1600 or 1150 cal BC temporal boundaries were deemed transitional.
This paper recognises the issue raised by Snoeck (2015) that most cremation burials are likely to suffer from the old wood effect to some degree.This would suggest that some burials whose date crosses the Early and Middle Bronze Age might in fact be Middle Bronze Age.A methodology for identifying the extent of this effect, or how best to treat cremated remains in this light, has yet to be produced.As a result only selected cremation burials were reclassified as Middle Bronze Age when their dates were only slightly older than 1600 BC and when there was evidence for other Middle Bronze Age cremation burials on the site.
Typological dating: Whilst typo-chronological schemes in metalwork, with a resolution of 150-200 years, are well established throughout the Bronze Age in Britain (Needham 1996;Needham et al. 1997;Roberts et. al. 2013), there are very few bronze or gold objects that have been found in secure and well excavated contexts with cremation burials.The vast majority of the cremation burials that are typologically dated to the Middle Bronze Age) rely upon associated ceramics.However, ceramic typochronologies typically have a more extended temporal resolution.For example, Collared Urns (c.1850-1500 cal BC) and Cordoned Urns (c.1900-1550 cal BC) are made, used, and deposited in the first century of the Middle Bronze Age, but are predominantly Early Bronze Age types (c.2200-1600 cal BC) (see Sheridan 2003;2007;Brindley 2007).Biconical Urns (c.1800-1400 cal BC) are more evenly divided across the Early-Middle Bronze divide (Tomalin 1988), whilst Deverel-Rimbury (c.1700-1200 cal BC) and regional variants such as East Anglian Ardleigh urns are predominantly, but not exclusively, Middle Bronze Age in date (Needham 1996, 132-3;Woodward 2009, 265-70).However, it is recognised that the majority of Deverel-Rimbury ceramics can be placed in the Middle Bronze Age (Needham 1996, 132-3;Woodward 2009, 265-70).A major challenge in chronological attribution has been the frequent revision of ceramic typochronologies as well as a wide variation in their adoption by key scholars.For instance, Grinsell's (1959;1971;1987;1992;O'Neil & Grinsell 1960) numerous corpora of barrows published by county, which have been exceptionally useful for identifying cremation burial sites, frequently mention 'Late Bronze Age' pottery types, many of which would now be typo-chronologically dated to the Middle Bronze Age.Similarly, records published prior to the use of radiocarbon dating in the mid-20th century which described 'Middle Bronze Age' pottery would now be placed within the Early-Middle Bronze Age range.In order to overcome this, typo-chronological adjustments had to be made on a site-by-site basis.To be fully transparent, all burials that have been judged to belong to a particular period include a reference that explains the reason they have been grouped this way in the Supporting Material (see column 'Reason-ForDate' in Appx S.1.2).

Fig
Fig. 1.The size of the 378 cremation burials sites when considering all cremation burials found on the site compared to the size of the 378 cremation burials sites when considering only those cremation burials that could be confidently dated to the Middle Bronze Age

Fig. 2 .
Fig. 2. Distribution of the 378 Middle Bronze Age cremation burial sites included in this study

Fig. 8 .
Fig. 8. Number of confirmed Middle Bronze Age cremation burial sites by distance to their closest potential Bronze Age settlement site.NB: the broken scale accounts for the high number of burials that were recorded as being in or around a settlement (the Y axis has been capped at 20 to better depict the distribution of sites, 49 sites were found less than 50 m distant from a settlement).

TABLE 1 :
STATISTICS DEMONSTRATING THE SKEW OF THE MINIMUM NUMBER OF BURIALS ACROSS MBA CREMATION BURIAL SITES WHEN CONSIDERING BURIALS OF ALL PERIODS & BURIALS SOLELY FROM THE MBA (378 SITES)

TABLE 2 :
COUNT OF MINIMUM NUMBER OF INDIVIDUALS (MNI) PER MBA CREMATION BURIAL (3133 BURIALS)

TABLE 3 :
MBA CREMATION CEMETERY SITES WITH MORE THAN TWO RADIOCARBON DATED CREMATION BURIALS Minimum and maximum durations have been calculated using the earliest and latest radiocarbon dates (at 95.4% probability) from cremation burials on the site.Key: *Those sites which have had Bayesian modelling.**Duetotheir proximity, the three sites at Bromfield (Bromfield Quarry, Bromfield Cemetery C1, & Bromfield Quarry Cemetery C3) and the four sites at Simons Ground (Sites B, C, F, & G) have been grouped together for the purpose of this table.***ManorFarm,Borwick was not studied, as it contained no radiocarbon dates placing a burial in the Middle Bronze Age.ADS = Archaeological Data Service.E.Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c.1600-1150 CAL BC)

TABLE 4 :
BREAKDOWN OF THE 3133 MBA CREMATION BURIALS WITH AND WITHOUT GRAVE GOODS Number of grave goods of each type found in the 3133 Middle Bronze Age cremation burials included in this study E. Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c.1600-1150 CAL BC)

TABLE 5 :
E.Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS,BRITAIN, MBA (c.1600-1150 CAL BC)funerary remainsin stark contrast to monuments and grave goods which characterise the archaeologically visible Early Bronze Age (c.2200-1600 cal BC) evidence (see Woodward 2000; Garwood 2007; Needham 2011; Wilkin 2011; Fowler 2013; Melton et al. 2013; Hunter & Woodward 2015; Jones 2016; Nicolas 2017).Where it exists, the evidence for Middle Bronze Age barrow construction is far less substantial in scale, investment, and complexity (eg, Bradley & Fraser 2010).Similarly, Middle Bronze Age personal ornaments were not placed with the dead as grave goods, but rather were worn and subsequently removed from the body and placed unburnt elsewhere (see Roberts 2007; Davies 2012; Wilkin 2017; O'Connor et al. 2017).There are only three poorly excavated Middle Bronze Age cremation sites where bronze ornaments have been found potentially in association(Roberts 2007, 149).It is certainly possible that ornaments could have been removed from the body as part of the funerary process.However, as particularly large ornament hoards, such as at Wylye, (Smith 1959;entifying the Wilkin 2017;O'Connor et al. 2017) BritainThe database assembled for this paper recorded 378 cremation sites containing at least 3133 burials that can be confidently dated to the Middle Bronze Age (c.1600-1150 cal BC).It is a very near comprehensive corpus of this feature type and represents the majority of archaeologically visible funerary evidence for the period.The analysis of the Middle Bronze Age cremation sites and burials demonstrates no bias towards particular sexes or age ranges.Neither is any social, ritual, nor political differentiation made explicit in the Wiltshire; West Ashling, Sussex; and Monkswood, Somerset(Smith 1959; Roberts 2007;Wilkin 2017;O'Connor et al. 2017), demonstrate, there is no straightforward equation between ornament hoards and individuals.
How comparable is the funerary record in North-west Europe c.1600-1150 cal BC?There are currently no in-depth comparative analyses of Middle Bronze Age funerary practices across North-west Europeor even between communities across the Irish Sea, North Sea, or Channel/Manche.
E.Caswell & B. Roberts.COMMUNITY CEMETERIES: CREMATION BURIALS, BRITAIN, MBA (c.1600-1150 CAL BC)mated human bone cannot have been used solely for this purpose, and probably was not in the majority of circumstances.De Mulder et al. 2014).As such, the radiocarbon dating of the remaining sites would assist in establishing whether certain absences in the funerary record, such as the northern England or East Sussex region, are reflective of the past reality.