Axes to Axes: the Chronology, Distribution and Composition of Recent Bronze Age Hoards from Britain and Northern Ireland

This study explores the impact that recent Bronze Age hoard finds have had on our understanding of hoarding practices across Britain and Northern Ireland. Changes to the legislation of Treasure and the onset of the Portable Antiquities Scheme in England and Wales have produced a wealth of new information on Bronze Age hoards. Beyond a handful of studies which have focused on specific groups of hoards or the distinction between dryland/wetland deposition, however, many of these more recent finds have been overlooked. Our regional understanding of hoarding practices across Britain is also largely based on studies which are now significantly out of date. This paper aims to address this problem by providing a snapshot of hoards and hoarding practices, based on a substantial dataset of 385 hoards (containing 7210 objects) that were reported on between 1997 and 2021. Broad chronological and spatial trends in the distribution are highlighted, with precedence given to characterising these enigmatic deposits based on their size and the categories of objects within them. This investigation provides fresh insights into the selection of certain object groups – particularly axes – during certain periods and within specific regions, whilst also exploring ideas so that we might better understand the scale of metalwork deposition. This research not only demonstrates how recent hoard finds fit into traditional narratives but also how they have the potential to enhance our understanding of regional hoarding practices, offering new and exciting avenues for future research.

For over 150 years, Bronze Age hoards have been a focus of interest for archaeologists.Early scholars valued them for their ability to tell us what kinds of objects were in circulation at a given time, forming the primary basis of the three-age system and the development of metalwork typologies from the mid-19th century (eg, Evans 1881).While some saw little value for hoards beyond their ability to construct typo-chronologies (Childe 1930, 44), others considered them to be an important source of information on life in the Bronze Age, with particular interest directed towards interpreting why hoards were buried.At one end of the debate there is an understanding that hoards were buried for safe-keeping and meant to be recovered; at the opposite end, is the belief that hoards were buried as ritual deposits that were never meant to be recovered.
Of course, the history of this debate and the ways in which ritual and utilitarian explanations have been made acceptable is complicated (see, for example Bradley 1998;Fontijn 2002).With the exception of some 19th and early 20th century scholars (eg, Evans 1881), archaeologists rarely surrender themselves completely to one interpretation or the other, instead arguing for the existence of both ritual and utilitarian hoards (eg, Levy 1982).While allowing for some degree of flexibility, this dichotomous approach remains problematic.Discriminating between ritual and utilitarian deposits assumes that one is exclusive from the other, a potentially anachronistic approach which overlooks the complexity of those patterns which we are able to observe (Brück 1999).In recent decades, the works of several authors have highlighted the immensely varied character of hoards and other metalwork deposits, arguing against the need to impose uniformity on past practices (Becker 2013;Bradley 2017;Fontijn 2002;2020;Needham 2001;2017).Instead, they have advocated for a data-led approach, returning to the empirical evidence as a starting point so that we can gain a more thorough understanding of the archaeological record and, in turn, practices of deposition.Moreover, what many of these recent studies have in common is an understanding that hoards are just one element of a 'spectrum of depositional practices', to borrow a term used by Cooper and her co-authors (2020;2022).As well as larger bodies of material, such as scatters of metalwork from settlement sites or grave goods, single finds are now more widely understood to also represent products of deliberate deposition, displaying similar depositional patterns to hoards (Becker 2013;Knight 2022, 121-4).The idea that single finds of metalwork could have formed the starting (or the end) point of formation for hoards has also been explored (Needham 2001;2017).
In recent decades, several studies have been carried out which have investigated specific groups of hoards (Mörtz 2018;Wiseman 2018), the distinction between dryland/wetland deposition (Yates & Bradley 2010a;Dunkin et al. 2020) or significant individual finds (Bradley et al. 2015;Adams 2017).A lack of recent and detailed empirical enquiry into the deposition of Bronze Age metalwork across Britain means, however, that our understanding of these practices remains vague.Only south-western England (Knight et al. 2015;Knight 2018;2022) has received a recently updated and comprehensive review, while the siting of Bronze Age hoards and single finds have also been explored in the context of north-east (Poyer 2015) and south-east England (Yates & Bradley 2010b;Dunkin et al. 2020).Recent re-appraisals of two metalworking 'traditions'assemblages which give a sense of unity, often involving a defined group of objects and materialshave also been undertaken.The first of these is the purported Middle Bronze Age 'Ornament Horizon(s)', a term which defines the practice of depositing copper-alloy ornaments within Taunton period hoards (c.1400-1275 BC), concentrated in central southern England, and the gold ornaments of the following Penard period hoards (c. 1275(c. -1150 BC) BC), which are more widespread in their distribution (Smith 1959;Roberts 2007;Needham 2017;O'Connor et al. 2017;Wilkin 2017).The second tradition is the Carp's Tongue/Boughton-Vénat Complex (Brandherm & Moskal-del Hoyo 2014, 24), a prominent group of hoards which are found across much of south-eastern England and northern France, whose deposition has been suggested to span across the closing century of the Late Bronze Age (c.900/875-800/775 BC) (Burgess 1968;Turner 2010;Brandheim & Moskaldel Hoyo 2014).
Such studies are significant for having much wider implications for the circulation of relevant object groups outside of their core distribution areas but a focus on these particular hoarding practices has meant that there has been relatively little consideration of regions outside of Wessex, the Thames Valley, and south-east England.Instances of other hoarding 'traditions' might be harder to detect but this does not negate the risk of 'national' narratives being established which are heavily reliant on evidence from a relatively small area (eg, Bradley 2009, 233-43).The above situation is particularly acute in Wales, despite parts of the country long being recognised as important sources of copper and, potentially, gold during parts of the Bronze Age (Northover 1995;Timberlake 2003;Williams & Le Carlier de Veslud 2019).Important contributions towards our understanding of the character and makeup of hoards from Wales were made over the latter half of the 20th century (Savory 1958;1980;Burgess 1968;Burgess et al. 1972;Needham 1981), many of which are now several decades old, resulting in a situation where our perceptions of hoards and depositional practices are heavily reliant on studies which are now significantly out of date.
In recent decades, significant work has also been undertaken on the development of object chronologies (eg, Davis 2012;2015;Brandherm & Moskal-del Hoyo 2014) and our overall understanding of the chronology of the Bronze Age (Needham 1996;Needham et al. 1997;Roberts et al. 2013).Of particular relevance here, however, are the significant numbers of hoard finds which have been reported from Britain and Northern Ireland over the past two decades.The majority of these new discoveries are from England and Wales, due, in large part, to the introduction of the Treasure Act 1996, its revision in 2002, and the success of the PAS (Portable Antiquities Scheme)details of both are described below.Wiseman's (2018) study highlighted the research potential of recent hoard finds from England and Wales but a focus on fragmentation THE PREHISTORIC SOCIETY patterns in 'scrap hoards' has meant that much of this material has been left without characterisation.
This article seeks to address some of these gaps in our knowledge, examining more closely the spatial and chronological patterns of hoards from across Britain and Northern Ireland, as well as outlining the number and categories of objects within them.The relevance of these recent hoard finds to previous arguments is explored, before concluding with some productive avenues for future research.

METHODS
This study covers the period c. 2200-800 BC, subdivided into the Early (2200/2150-1550 BC), Middle (1550-1150 BC), and Late (1150-800 BC) Bronze Age, presenting a long-term overview of recent hoard finds from Britain and Northern Ireland.The British Chalcolithic (c.2450-2200/2150 BC) and the Earliest Iron Age (c.800-600 BC) are sometimes included within syntheses of the Bronze Age (eg, Poyer 2015; Knight 2022), but are both excluded here.In Ireland, the metalwork chronology uses different names and date ranges for each of its stages, although they are broadly comparable to those used for the Bronze Age in Britain (Eogan 1983;Waddell 2000;Becker 2012;Roberts et al. 2013).Where appropriate, the Irish Metalworking Assemblages have been detailed alongside their corresponding British (ie, mainland) Metalworking Assemblages (eg, Table 2

below).
The development of radiocarbon dating techniques and the increased number of independent dates available for artefacts has resulted in a more refined typochronology specific to British metalwork (Needham 1996;Needham et al. 1997;Roberts et al. 2013), which the chronology presented here follows (Table 2, below).The reason why there are some gaps or overlaps between some Metalworking Assemblages (eg, Taunton and Penard) is not solely because of issues surrounding the refinement of dating evidence but because they are based on interlinking associations between certain object groups rather than a rigid temporal sequence (Needham 1996, 123;2017).As described by Needham, Assemblages are essentially chronological 'behaviour packages', as much about rules which governed processes of circulation and deposition as they are about chronological or production factors (Needham 2017, 130, 151).Even though they themselves have temporal limits, the boundaries of Assemblages may range from sharp to diffuse depending on whether key combinations change suddenly or gradually from one mode to another (Needham 2017, 113).
In total, 385 Bronze Age hoards, reported on during the period 1997-2021, are considered within this paper, containing approximately 7210 objects (7170 metal and 40 non-metal).Within this study, hoards are defined as two or more closely associated precious or base-metal objects that derive from a single deposit (ie, adopting the definition currently used for Treasure in England and Wales).This definition also includes any non-metal artefacts found in association with groups of metal objects.Secondly, it includes scattered groups of objects once probably buried in direct association but disturbed and scattered in more recent times via secondary processes (eg, ploughing). 1 The data was compiled from published and unpublished sources and catalogues, the full details of which are described in the Appendix S1.An additional 14 finds were considered as possible hoards but have been excluded from all following analyses.The details for these 'possible hoards' and the reason(s) for their exclusion are included in Appendix S2.
A large proportion of the hoards under consideration within this study (364 of 385) qualified as 'treasure' at the time of their discovery, falling under the Treasure Act 1996 and the Treasure (Designation) Order 2002, as well as the Scottish Law of Treasure Trove.Enacted in 1997, the Treasure Act 1996 replaced the previous Treasure Trove law of England, Northern Ireland, and Wales, removing the need to demonstrate the motivations of the depositors (and the inferred intention to return to retrieve objects) as integral to demonstrating whether objects were Crown property or could be kept by the finder, and making it mandatory to report discoveries of gold and silver objects older than 300 years.The Treasure (Designation) Order 2002 extended this definition to include finds of two or more base metal (ie, any metal other than gold or silver) prehistoric objects, meaning that all Bronze Age metalwork hoards discovered since the beginning of 2003 now qualify as treasure.The impact that the Order has had on the reporting of Bronze Age hoards is significant.Of the 367 hoards included within this dataset from England, Northern Ireland and Wales, 345 were discovered and qualified as Treasure between 2003 and 2021, compared with the five hoards which contained precious metal and were declared as treasure between 1997 and 2002.The remaining 17 base-metal hoards from England C.J. Griffiths.CHRONOLOGY, DISTRIBUTION & COMPOSITION, RECENT BRONZE AGE HOARDS, BRITAIN & N. IRELAND and Wales did not qualify as Treasure at the time of their discovery but were reported on and published by the PAS.
The wide geographical coverage of this study means that it is important to acknowledge the differences between jurisdictions, particularly those laws which concern 'treasure' and metal-detecting.Failure to consider the impact of these modern practices runs the risk of reporting on patterns which might not be reflective of genuine Bronze Age hoarding activity.For example, the Treasure Trove system in Scotland is distinct from England, Northern Ireland, and Wales as all portable antiquities of archaeological significance, regardless of the material they are made from, must be reported and are subject to claim by the Crown (KLTR 2016).Eighteen Bronze Age hoards from Scotland are included within this dataset, a relatively small total compared to those of England and Wales.Some authors (eg, Saville 2009, 95-6;Dalton 2014) have previously argued for significant levels of underreporting among metal detectorists in Scotland, the reasons for which are likely to be complex and not necessarily a case of wilful non-reporting of finds (see Dalton 2014 for a good summary of the debate).In Northern Ireland, The Historic Monuments and Archaeological Objects (NI) Order 1995 stipulates that it is an offence to search for and remove archaeological objects without a 'license to excavate' issued under the Order.Considering that most hoard finds are, these days, discovered by metaldetectorists (cf.Murgia et al. 2014, 358), it is highly probable that the low number of finds from Northern Ireland presented within this paper are more a reflection of the strict limitations on the use of a metal detector, rather than a reflection of genuine hoarding practices.
Although the research potential of the finds recorded by the PAS is significant, the scheme is not an unbiased source of evidencean aspect which the PAS has itself been keen to stress (Robbins 2014).Of the many factors identified by Robbins, those which are likely to have the biggest impact on this dataset are areas with strict constraints on metal detecting (eg, Forestry Commission land, Scheduled Ancient Monuments, National Trust land, or military zones), as these are areas where it illegal to use and be in the possession of a metal detector without permission from the appropriate authority.Extensive development within urban areas also makes it difficult for metal-detecting but other factors such as land elevation can also be a significant deterrent.For example, Robbins' (2014) study identified that only 1.9% of PAS finds reported up to 2013 were recovered from above the 200 m elevation line, perhaps somewhat accounting for the low numbers of Bronze Age hoards which have been recovered from certain areas during the last 25 years, such as central Wales and parts of northern Britain.Other biases in the identification and recording of artefacts, perhaps influenced by different research interests or perceptions amongst those recording finds (Robbins 2014, 36), are unlikely to have a significant impact here, especially considering that Bronze Age hoards are legally considered Treasure, with a handful of specialists handling the reporting of these cases within their respective countries.PRE-1997VS POST-1997 This paper primarily deals with those hoards which were reported on between 1997 and 2021 but it is important to consider how these recent hoard finds relate to those which were discovered prior to 1997.To put the 385 hoards under detailed consideration here into perspective, approximately 1100 hoards are reported as being recorded on the card index of Bronze Age finds held at the British Museum, which lists finds made up to 1985, whilst approximately 200 hoards were estimated as being discovered between 1985 and 2003 (Bland 2015, 2).Accounting for a slight overlap with Bland's estimate for the period 1985-2003 and excluding the three hoards from Northern Ireland, the dataset considered here is estimated as representing approximately 23% of all known Bronze Age hoards reported from Britain.
Whilst the numbers presented above offer some insight into the possible shortcomings of this dataset, they reveal little about the impact of recent finds on a regional level.For example, Coles listed five Middle Bronze Age (1963) and over 60 Late Bronze Age hoards (1960) from Scotland, the latter of which are concentrated across the south and north-east of the country (Coles 1963, appx 3).Compared with the historic Late Bronze Age hoard record, the relatively few recent finds from north-east Scotland are, perhaps, suggestive of a reporting/recovery bias within this part of the country.For Northern Ireland, Becker's (2006) study highlights the strength of the pre-1997 hoard record, curtailing most meaningful observations which might be made about this region from the post-1997 dataset.
For England and Wales, where the majority of recent discoveries have been made, more detailed regional comparisons between pre-and post-1997 hoard THE PREHISTORIC SOCIETY finds are possible.To date, there are no published datasets which might be drawn upon to offer a complete picture of the quantities and distribution of pre-1997 hoard finds.Studies by Rowlands (1976) and Huth (1997) have previously been used, in combination, to compare the distribution of historic and more recent Middle and Late Bronze Age hoard finds (Wiseman 2018, 40, fig.1), but neither study can be considered to offer a complete overview of their respective sub-periods.Both studies exclude Wales and northern England, whilst an emphasis on Rowlands' (1976) study would result in over two decades of Middle Bronze Age hoard finds from southern Britain being unaccounted for.
Considering that a full synthesis of Bronze Age hoards would be a significant undertaking and beyond the scope of this project, it was decided to focus on seven case study regions for detailed data collection to facilitate comparison between pre-and post-1997 finds (Figs 1-2, Table 1).South-east and west Wales were chosen because this paper stems from a project investigating Middle and Late Bronze Age hoards and hoarding practices from both regions, whilst access to additional unpublished works (eg, Northover n.d.) facilitated the decision to also include central and north Wales and The Marches.Outside of Wales and The Marches, other regions selected were those where recent and comprehensive studies of Bronze Age metalwork have been undertaken: south-west England (Pearce 1983;Knight et al. 2015;Knight 2018;2022), northeast England (Poyer 2015), and East Anglia (Pendleton 1999).Although there are gaps between these study regions, their spatially dispersed nature provides a good balance between south and north, lowland and upland, coastal and inland.
As demonstrated in Figure 1 and Table 1, the relationship between the pre-and post-1997 datasets is complicated.For example, 56% (n=47 of 84) and 47% (n=9 of 19) of all Bronze Age hoards reported from, respectively, south-east and west Wales were discovered during the period 1997-2021.By contrast, only 12% (n=9 of 76) of all Bronze Age hoards reported from central and north Wales were discovered between 1997-2021.Of particular interest is the observation that, with the exception of central and north Wales, the number of Late Bronze Age hoard discoveries for all case study regions appears to be disproportionately weighted towards the post-1997 dataset (ie, significantly more Late Bronze Age hoards have been found per year since 1997).Furthermore, based on the results of the present author's detailed investigation within areas with high quality, accessible datasets (as outlined above), the percentage of post-1997 hoard finds appears to represent on average 28% (n= 198 of 710) of all Bronze Age hoard finds in these case study regions.As well as accentuating those patterns described above, Figure 2 demonstrates that there is a fairly good match between the distribution of pre-and post-1997 hoard finds, particularly from south-east and west Wales, Norfolk, Suffolk and, to a lesser degree, north-east England.The relatively few recent hoard finds from central and north Wales have already been mentioned but parts of south-western England and Cambridgeshire also appear to be poorly represented by the post-1997 datasetregions which are, historically, well-represented by Bronze Age hoard finds.The reasons for these disparities are unlikely to be straightforward (see Robbins 2014; Cooper & Green 2017), and it is not the intention of this paper to become caught up in such a discussion.Of course, focusing on post-1997 hoard finds will always mean that we are dealing with a significant sample rather than a complete dataset but the benefits of focusing on recent hoard finds far outweighs the negatives.The PAS web database, in particular, contains a vast quantity of high quality data (eg, object types, findspot location, known contextual details) which are easily available and free to access digitally (https://finds.org.uk/database).Also, under the collaborative PAS system in England and Wales, metal-detectorists are now generally more informed about the potential significance of small objects or fragments.This is particularly significant with regards to Late Bronze Age hoards, as these tend to consist of multiple, small fragments of bronze which could easily be overlooked or discarded.That finders are now legally compelled to declare prehistoric base-metal hoards (as well as being rewarded) also means that greater attention is paid to the provenance of finds, whilst the increasing involvement of archaeologists in the excavation of hoard finds reveals substantially more information about the context in which the objects were buried.In addition to these benefits, the understanding that more recent finds account for 23-28% of all Bronze Age hoards (see above) means that we can be more confident in their ability to contribute constructively towards our understanding of broader scale processes, to which we now turn our attention.

CHRONOLOGICAL AND SPATIAL HOARDING PATTERNS
The number of hoards and artefacts represented within this dataset are detailed in Table 2 and depicted in Figure 3, clearly demonstrating considerable variety throughout the Bronze Age.Relatively few hoards were deposited during the Early Bronze Age and they typically contain small numbers of objects.A noticeable increase in both the number of hoards and the objects deposited within them is visible in the Taunton phase of the Middle Bronze Age, before dropping in the Penard phase.Relatively low numbers of hoards are maintained into the beginning of the Late Bronze Age, though the number of objects rises dramatically thereafter.We then see the highest peak in both the number of hoards and objects within them towards the end of the Late Bronze Age, during the Ewart Park phase, when the number of hoards being deposited is significantly higher than that of the previous 300 years combined.
The patterns outlined above are a very broad overview, masking potentially more variable regional temporal trends.Figure 4 depicts the distribution of   5.
Of relevance here is the possibility that these observed hoard frequencies might simply reflect modern biases and variations in metal-detecting activity and reporting.With this in mind, Figure 6 compares the total number of finds recorded through the PAS (Neolithic to modern) with the number of Middle and Late Bronze Age hoards reported, for ten case study regions.These regions were selected based on the availability of data as well as their relevance to those spatial patterns described above, and to the parent project which focuses specifically on south-east and west Wales.Some regions (eg, Wiltshire and Gloucestershire) have been combined so that they are broadly similar in area (km 2 ).All percentages in this section are expressed in relation to the number of Middle (n=53) and Late Bronze Age (n=151) hoards from the ten case study regions.To a considerable extent, Figure 6 indicates that the frequency of hoard discoveries is not necessarily a reflection of the intensity of metal-detecting or reporting  does not appear to have had much of an impact on the number of hoards reported from both regions.Figure 6 also has the benefit of emphasising some of those chronologically specific patterns described above.Most notably, Wiltshire/Gloucestershire and Somerset/North Somerset/Dorset have relatively high numbers of Middle Bronze Age hoards (seven and ten respectively) which contrasts strongly with the proportionally low number of Late Bronze Age hoards from both regions (four and five respectively), suggesting that these frequencies are reflective of genuine later prehistoric hoarding activity and not modern non-archaeological practices.The distribution of Early Bronze Age hoards, which are typically far less frequent, in regions where Middle or Late Bronze Age hoards are relatively rare (see Fig. 3), also offers further evidence to support such observations.

SIZE OF HOARDS
Further meaningful observations can be made about the general character and composition of Bronze Age hoards.For example, the size of hoards can be used to gain further insight into regional hoarding practices, particularly for the Middle and Late Bronze Ages when the majority of copper-alloy and gold objects were deposited.The size of hoards can be measured in two ways, by weight or by the number of objects, and choosing one over the other could result in different implications.For example, palstaves are typically heavier than socketed axes and this might offer up an interesting comparison between the weight of bronze being deposited in certain areas and/or during certain sub-periods of the Bronze Age.Weights of hoards or individual objects are not, however, detailed on those records which were accessed for this study and so the decision was made to focus on the number of objects within hoards.Worth noting is that, although there is an emphasis here on the number of individual artefacts within a hoard, it is not always possible to distinguish this from the number of individual fragments, especially for many Late Bronze Age hoards.For example, the Tattershall hoard from near Stixwould, Lincolnshire (PAS LIN-CEDC78; Treasure Case 2006 T308) contains 161 copper-alloy fragments, some of which possibly derive from the same individual object.In these cases, the minimum number of metal objects In Figure 7, the number of metal objects within hoards is summarised for the Early, Middle, and Late Bronze Ages.To a considerable extent, these results map closely onto the data displayed in Table 2 and  Figure

OBJECTS WITHIN HOARDS
Bronze Age hoards contained a wide variety of objects, particularly those deposited during the Middle and Late Bronze Age.There are different ways to break this data down to look for regionally or chronologically specific trendsone is to assign hoards a category based on the types of objects that predominate.These categories and their relevant object types are summarised in Table 3.
For a find to be categorised as an 'ornament hoard', for example, it would need over half of its composition to be made up of objects related to personal adornment.'Mixed hoards' are used to describe hoard finds where no single category of object makes up more than half of its composition, such as the Llanfrynach Community hoard (NMGW-47D2BB; Treasure Case Wales 16.18) which contains a minimum of ten (possibly 11) objects, including at least two socketed axes, fragments representing at least two bracelets, a pin head fragment, an annular ring, a spearhead tip fragment, a sword blade fragment, and two casting jets.Hoards which contain approximately equal quantities of objects from two object categories are also treated as 'mixed hoards'; for example, the Taunton-Penard phase hoard from East Peckham, Kent (KENT-20E688; Treasure Case 2021 T631), which contains one rapier blade fragment and an incomplete copper-alloy pin.Previous studies have sometimes grouped axes and tools together (cf.Dunkin et al. 2020, 74) but the abundance of axes within Bronze Age hoardscompared with the common, but relatively minor, occurrence of knives, chisels, and gouges, etcwarrants them being considered as a separate category.Figure 9 depicts the prevalence of hoard categories, as described above, through time, according to each of the Metalworking Assemblages used within this study.Axe hoards are the most prevalent category throughout the Bronze Age, accounting for almost all Early Bronze Age hoards within this dataset (n=15 of 16), 44% of all Middle Bronze Age hoards (n=34 of 77), and 34% of all Late Bronze Age hoards (n=97 of 291).The Taunton and Penard phases of the Middle Bronze Age stand out for the relatively high proportion of ornament hoards, representing 30% of all hoards from this period (n=20 of 67).Eleven ornament hoards are also present during the Late Bronze Age but the lack of associated radiocarbon dates means that most of these have been given an 'unphased Late Bronze Age' date.Two more precisely dated examples are included here also, including the Ewart Park phase hoard from North Cove, Suffolk (SF-BDA986; Treasure Case 2011 T478), where a ribbed socketed axe of Class B (Southern English type) was found with five gold lock-rings, all contained within the socket of the axe.The proportion of metallurgical hoards jumps drastically during the Late Bronze Age, accounting for 26% of those deposited between 1150 and 800 BC (n=76 of 291) compared with 5% for the preceding Middle Bronze Age (n=4 of 77).
With the exception of mixed hoardswhich are explored below -Figures 10 and 11 depict the distribution of, respectively, Middle and Late Bronze Age hoards based on those categories which were outlined above.As well as providing a means of exploring regional trends it is possible to refine this broad-brush approach further by distinguishing hoards with a dominant object category (such as axe dominant hoards) from those which only contain objects from one category (such as weapon only hoards).In the Middle Bronze Age, central southern England stands out for its concentration of ornament hoards, the majority of which correspond with the Taunton phase (cf.Roberts 2007).Eleven of the 18 ornament hoards from southern England only contain objects of personal adornment, with bracelets and   Riding of Yorkshire (PAS YORYM-D2333A; Treasure Case 2016 T240); it contains a total of 150 objects including a substantial quantity (91) of ingot and ingot fragments along with a range of complete and fragmented socketed axeheads (Huisman & Haldenby 2018).Most, if not all, these ingot fragments are of plano-convex type, which are a frequent inclusion in late Ewart Park hoards from south-eastern England and western France (Le Carlier de Veslud et al. 2013, 509-10).
The variety of content within mixed hoards warrants further elaboration.Figures 12 and 13 exclusively depict the distribution of mixed hoards, labelled according to the categories of objects within them.It is important to note that, in presenting this data, these labels do not provide a proportional representation of the kinds of objects within hoards (as this would have been too complicated to depict) but rather simple presence/absence.Nevertheless, Figure 12 compliments some of the patterns observed above, such as the frequent occurrence of ornaments within Middle Bronze Age hoards across central southern England.Depicting the composition of hoards in this manner also has the benefit of highlighting additional aspects of this dataset.Most mixed hoards of the Middle Bronze Age contain two (n=7) or three (n=4) object categories, with the two most 'complex' hoards each containing four object categories.Axesspecifically palstavesare unsurprisingly the most common object within mixed hoards of this period (n=10), followed by weapons (n=9), ornaments (n=7), tools, and metallurgical material (both n=4).The relatively common occurrence of weapons within mixed hoards contrasts with the situation outlined earlier, in Figure 9, where weapon hoards are relatively rare.This suggests that, on the rare occasion when weapons were placed exclusively together, this may have been associated with more particular social or cultural connotations.
Figure 13 offers a more complex and difficult to untangle picture of mixed hoards which date to the Late Bronze Age.Out of the 85 mixed hoards included here, the majority contain three (n=25) or four (n=24) object categories, indicating that mixed hoards of this period are generally more varied in their composition than those deposited during the Middle Bronze Age.It is also during this period that the most complex hoards first appear.Three hoards contain all six object categories adopted for use within this study: the previously mentioned Boughton Malherbe hoard, the Crundale hoard, which is also from Kent (KENT-7C3863; Treasure Case 2003 T374), and the Barton Turf CP hoard, Norfolk (NMS-6DAAFAC; Treasure Case 2016 T470).Axes are once again the most dominant object category, generally, within Late Bronze Age mixed hoards (n=71), followed closely by metallurgical material (n=64) and weapons (n=63).More notable are the relatively high proportions of hoards containing tools (n=51) and material classed as 'other' (n=34), contrasting strongly with the underwhelming proportion of tool and 'other' hoards depicted in Figures 9 and 11.DISCUSSION This is the first study to compile data on Bronze Age hoards at this scale, allowing for a broad scale visualisation and analysis, on a firm empirical basis, of temporal and spatial trends in hoarding practicesparticularly in Britain.Through large scale data collection, it has been possible to trace the sudden increase in hoarding deposition from c. 1400 BC across parts of southern England, the drop in the succeeding Penard and Wilburton phases, and the abundance of Ewart Park hoards relative to earlier periods in many regions.These observations will, of course, come as no surprise to those with prior knowledge of Bronze Age hoarding practices, particularly the peak in deposition during the Ewart Park phase which has been used to infer aspects of the role of metal, either as caches of material for recycling or as evidence for a collapsing bronze economy (Burgess 1979;Thomas 1989;Wiseman 2018).Other well-known patterns, such as the Middle Bronze Age 'Ornament Horizon(s)' and the Late Bronze Age Carp's Tongue/Boughton-Vénat Complex are also well represented by recent hoard finds; the former, by the distribution of ornament dominant/only hoards across central southern England and the latter by the concentration of large and mixed hoards across south-eastern England.Being able to prove these well-known trends not only validates previous studies and demonstrates the validity of this dataset but it is also extremely rewarding to now be able to visualise these patterns on the basis of a firm and strengthened empirical foundation.
Well-known patterns which have not always been attested empirically have also emerged throughout this study.For example, when compared with the evidence for Middle Bronze Age and Earliest Iron Age hoards ), the dearth of Late Bronze Age hoards across parts of south-western and central southern Englandparticularly over Somerset, north Dorset, and Wiltshireis now abundantly clear.Explaining why there was a decrease in the number of hoards deposited across these parts of southern England during the Late Bronze Age, a period when the rate of hoard deposition generally increases across most regions, particularly at 1000-800 BC, is more difficult and warrants consideration of contemporary social and economic contexts.Southern England has contributed vast quantities of evidence for an intensification of settlement and agriculture during the Middle and Late Bronze Age (eg, Yates 2007).Product surpluses, built from this intensification of farming, may have provided communities with the means to produce and acquire more metalwork, leading to, it might be argued, more prolific depositions of metalwork (Yates 2007, 119).When applying this model to those areas which have significantly fewer hoards during one period over another, we might therefore expect to see a parallel converse situation, with lower frequencies of settlement evidence and lack of intensification of agriculture.Compared with the Middle Bronze Age there is, however, no observed significant fall in the density of known settlement across those parts of south-western and central southern England where Late Bronze Age hoards are relatively uncommon (Caswell 2020, fig. 81).This suggests that there is no simple or direct correlation between an intensity of settlement and the frequency of hoard deposition.A similar conclusion might also be reached with regards to south-east Wales, where the sparsity of evidence for contemporary settlement is an important contradiction to the otherwise frequent deposition of Ewart Park hoards (cf.Burrow 2020).
Some other patterns that emerged out of this this data were far more surprising.For example, the dearth of large hoards from western Britain, Scotland, and Northern Ireland appears striking.However, it is crucial to understand that large hoards have historically been reported from these regions; for example, the Wilburton phase hoard from Guilsfield, Powys, or the Ewart Park hoards from Stogursey, Somerset, and St Andrews, Fife (Barnwell 1864;McNeil 1973;Cowie et al. 1998).One point springs to mind when looking at the difference in the size of deposits: the temptation to view these as measures of 'wealth'.A natural extension of those ideas expressed by Yates (2007, 119) might be to view those communities with fewer or smaller hoards as less prosperous or far removed from Late Bronze Age exchange networks, resulting in less prolific deposits of metalwork.The distributions of metallurgical dominant/only hoards along the southern coast of south-west England and west Wales offer an important contradiction to this idea, however, as the presence of copper/copper-alloy ingots implies the passage of raw material into or out of their respective regions.As noted by Knight with regards to recent hoard finds from Cornwall, the presence of objects most commonly associated with Carp's Tongue/Boughton-Vénat hoards suggests maritime exchange and influence from north-western France (2022,.A more nuanced understanding of the material makeup of Late Bronze Age hoards from west Wales remains to be demonstrated but it is significant that evidence for Carp's Tongue material has also been noted amongst several recent hoard finds from this region and south-east Wales (Gwilt et al. 2011;2014;Knight 2022, 104).What the material makeup of these hoards suggests is that communities along these south-west, sea facing regions were active participants in long distance cross-Channel exchange, implying that their contents were not entirely random.Furthermore, although the size of a hoard can give an impression of prosperity, there is no real reason why the amount of material being deposited should be viewed as directly proportionate to the wealth of individuals or communities.
With the above discussion in mind, the question must be asked: how might we recognise and interpret the meaning of hoards?It might be tempting to seek out alternative causation factors to explain chronological and regional variations in the hoarding record.For example, the concentration of Late Bronze Age ringworks in south-east England, particularly along the Thames Valley, has been characterised as representing high status enclosures which occupy strategic positions and correspond with concentrations of metalwork and field systems (Yates 2007, 24-6, 119;Brown & Medlycott 2013, 152-5;Evans et al. 2016, 214-16).The development of ringworks in south-east England may seem to be an important byproduct of cross-Channel exchange and relations during the Late Bronze Age but similar conclusions cannot be easily met for regions where evidence for similar enclosed settlements is lacking.
Thought needs to now be directed towards how we might best understand the scale of metalwork C.J. Griffiths.CHRONOLOGY, DISTRIBUTION & COMPOSITION, RECENT BRONZE AGE HOARDS, BRITAIN & N. IRELAND deposition in the Bronze Age.At their most fundamental level, hoards represent groups of metalwork which survive archaeologically, meaning that we must draw our conclusions from what was buried in the ground and never meant to be recovered, what was left after some material had been taken out again, and what may also have been lost, forgotten, or deemed insignificant.Considering that much metalwork was destined to be recycled or to continue in circulation (Needham 2001;Bray & Pollard 2012;Wiseman 2018), we must acknowledge that the hoard record, and the wider metalwork repertoire, represents the exception rather than the norm.This indirectly relates back to Needham's (2017) concept of Assemblages as a form of depositional phenomena, where our understanding of Bronze Age circulation, exchange, and deposition is chronologically and geographically varied as a result of human agency.More recently, Fontijn has argued for the existence of a broader set of values which guided the actions and motivations of individuals in Bronze Age Europe (2020,.This set of guiding principles applied not only to those objects which we might consider to be inalienable or of outstanding character, but also the seemingly alienable objects which may have been akin to commodities (Fontijn 2020, 44-60).The point here is that, at some stage, the social value of depositing and leaving objects in the ground was judged to outweigh any intrinsic value to be gained through their continued circulation or recovery.
The knowledge that, in general, Bronze Age hoards only comprise a selection of material once in use and circulation supports the notion that the types of objects buried were also subject to a guiding set of principles (Needham 1988;Fontijn 2002;2020;Becker 2013;Cooper et al. 2022, 75-110).Turning now towards selection of certain objects, we can be more confident in interpreting the contents of recent hoard finds, especially as most of the biases which affect this dataset are concerned with distribution patterns.In Bronze Age Britain, the persistence of axe hoards is especially noteworthy, particularly in the Late Bronze Age when the variety of metal objects produced, circulated, and deposited reached its peak.Of course, considering that axes took on a great variety of shapes, sizes, designs, and even colour during the 1400 years under consideration here, it may be misleading to label all these objects under a modern collective term.Nevertheless, it seems important that most objects buried within Bronze Age hoards had the shape of an axe (Barrett 1989, 315), raising the question of how the selection of axes relate to the broader system of value in the Bronze Age.
Fontijn has emphasised how the capacity for bronze to communicate value is very much based on its shape and that, in certain transactions, items ought to have prescribed and widely accepted shapes (Fontijn 2020, 86-111).In a shape-based value system which encompassed bronze objects of various forms, axes may have constituted a specific sphere of values which the act of deposition, and the performance around it, served to anchor socially (Fontijn 2020, 105).Whether flat or flanged axes, palstaves and socketed axes were necessarily understood as 'the same' is perhaps unlikely but this does not negate from the prominence of axes as a broad object category throughout the British Bronze Age.That there were sub-periods and regions where the deposition of axe and axe hoards appears to be less prominent, such as across parts of central southern England in the Middle Bronze Age, suggests that different social and ideological values were being expressed through hoarding practices.By the Late Bronze Age there appears to have been a greater degree of flexibility regarding what types and combinations of objects were deemed appropriate for deposition.The significant proportion of Ewart Park axe dominant/only hoards across south-east Wales is, however, perhaps suggestive of a strong tendency towards maintaining traditional depositional practices.The concentration of a distinct style of socketed axethe South Wales/ Stogursey Type (Burgess 1968;2012;Needham 1981;Gwilt 2004)within the region may not only reflect the popularity of a common, all-purpose woodworking tool, but perhaps also served to communicate certain ideas pertaining to regional identity or status.That similar ideas appear also to have existed during the Earliest Iron Age (cf.Boughton 2015) when, overall, much less metalwork entered the ground, strongly suggests that these ideas continued to hold true into the 8th and 7th centuries BC.

CONCLUSION -LOOKING FORWARD
As noted earlier, focusing on recent hoard finds does not tell us everything there is to know about Bronze Age depositional practicesthis cannot be overstressed.What this study does do, however, is demonstrate a means of marshalling, visualising, and understanding large scale and emerging trends THE PREHISTORIC SOCIETY in the data generated over the past 25 years through Treasure and the Portable Antiquities Scheme.The combined influence of these changes has previously been explored in relation to gold objects (Murgia et al. 2014), object types (Davis 2012;2015), and on a regional level (Knight et al. 2015;Poyer 2015), but this is the first study of its kind to combine all these factors together, whilst also presenting a crosschronological analysis of hoards throughout the Bronze Age.In so doing, it has proved possible to offer confirmation for several important and widely held beliefs regarding the deposition of hoards, bringing to the fore the growth in hoarding abundance over time, as well as important regional and chronological differences.This study has also highlighted how, at this broad a scale of analysis, it is possible to suggest that hoards were subject to specific selection.In particular, axe hoards are shown to be a significant element of the depositional material repertoire, offering insight into the existence of complex ideas attached to the significance of these artefacts.
At the end of this paper, it must be acknowledged that it has only been possible to scratch the surface of this extremely rich dataset.The nature of studies such as this is that they can also give the impression of highly synchronised spatial and temporal changes, which may very well have been the case, but we must allow for regional assemblages to speak for themselves.Of note are the recent hoard finds for southeast and west Wales, accounting for over half of the known corpus of material from both regions combined.A study and re-assessment of Middle and Late Bronze Age metalwork finds from both of these regions is already underway, including a complete review and synthesis of the South Wales/Stogursey Type socketed axe which features so heavily within Ewart Park hoards from south-east Wales.More focused work on object types will not only contribute towards our understanding of cross-Channel material mobility throughout the Bronze Age but also connections across land and the Irish Sea region.Analysis of patterns of wear on objects would not only glean further information on the social role of objects (cf. Fontijn 2002, 212), but would also help to identify those objects which may have been in circulation for extended periods of time.The question of how metalwork deposition relates to the wider deposition of non-metalwork also remains.The study by Cooper et al. of the 'ebb and flow' of objects deposited within hoards, burials, and settlement could be extended to other regions and sub-periods, although clearly much work remains to be done in bringing certain datasets up-to-date (2022, 92-3, table 4.02).In the future, both strands of evidence might be brought together to construct a more effective and holistic understanding of depositional processes in the prehistoric past.It is hoped that the evidence set created here will provide a foundation on which much of this work can be built upon, allowing for richer stories of objects and people in the Bronze Age to be told.

Fig. 1 .
Fig. 1.Relative proportion of pre-versus post-1997 Bronze Age hoard finds by case study region

Fig. 2 .
Fig. 2. Comparison between the distribution of pre-and post-1997 Bronze Age hoard finds by case study region

Fig. 3 .
Fig. 3. Total number of hoards, plotted alongside the total number of objects within hoards through time (per 50 year timeslice).Hoards of unphased Early, Middle, or Late Bronze Age have been excluded so as not to skew the results

Fig. 5 .
Fig. 5. Distribution and relative density per km 2 of Middle (left) and Late (right) Bronze Age hoards (contains OS data ©Crown copyright and database right 2022)

Fig. 6 .
Fig. 6.Relative proportion of archaeological objects (Neolithic-modern) recorded through the PAS and the number of Middle (MBA) and Late (LBA) Bronze Age hoards reported by case study region 3, confirming that the increased frequency of hoarding is synonymous with hoards generally becoming larger in size as the Bronze Age progresses.Almost 90% (n=14/16) of Early Bronze Age hoards contain 2-5 objects, with the largest hoard of this period being the previously mentioned Arreton (Metalwork Assemblage VI) phase hoard from Westenhanger, Kent, with 15 objects.Small hoards, which contain 2-5 and 6-10 objects, are also frequent in the Middle and Late Bronze Age.However, and large sized hoards (41-60 and 61≤ objects) become progressively more frequent, accounting for 18% (n=13) of all Middle Bronze Age hoards and 50% (n=124) of all Late Bronze Age hoards included within this dataset.The Taunton phase hoard from the Lewes Area, East Sussex (SUSS-C5D042; Treasure Case 2011 T192) is the largest Middle Bronze Age hoard, consisting of 79 objects including: 3 copper-alloy palstaves, 53 copper-alloy ornaments (complete examples and fragments), 4 sheet gold discs, and 19 amber beads.Worth highlighting is the Penard phase hoard from Cirencester, Gloucestershire (BM-28B710; Treasure Case 2004 T416), which contains the most individual metal objects including: 3 copper-alloy awls, 1 copperalloy knife, 1 copper-alloy spearhead, and 57 objects of goldmostly fragments of various forms of personal ornamentation.By comparison, the largest Late Bronze Age hoard within this dataset is the Wilburton phase (c.1150-1020 BC) Preston Hill hoard from Kent (PAS KENT-DA6E86; Treasure Case 2016 T450), where 929 fragments of copper-alloy plate were discovered within a ceramic vessel.Worth emphasising, however, is that the largest hoard by weight is the late Ewart Park hoard from Boughton Malherbe, Kent (PAS KENT-15A293; Treasure Case 2011 T464; Adams 2017), which contains c. 340 objects with a total weight of 64.2 kg, compared to the 12.5 kg of the Preston Hill hoardhighlighting the different ways in which hoard contents can be described and emphasised.2Figure8illustrates the geographical distribution of Middle and Late Bronze Age hoards according to their size.For the Middle Bronze Age, small hoards are distributed throughout southern England, East Anglia,

Fig. 7 .
Fig. 7. Numerical frequencies of hoard sizes during the Early, Middle, and Late Bronze Age

Fig. 8 .
Fig. 8. Distribution of Middle (left) and Late Bronze Age (right) hoards, according to their size (contains OS data ©Crown copyright and database right 2022) THE PREHISTORIC SOCIETY torcs made of both gold and copper-alloy being the most characteristic form of object within them.Axe hoards are the most widespread category and are found in all regions where Middle Bronze Age hoards have been discovered.The concentration of axe only hoards across coastal areas is also noteworthy, particularly in Kent, where eight out of the nine reported hoards contain only palstaves.As mentioned previously, metallurgical hoards represent a small minority of Middle Bronze Age hoards within this dataset, representing just four examples from the subperiod.The two metallurgical only hoards -Conwy Community, county of Conwy (Treasure Case 17.12 Wales) and Hempnall, Norfolk (SF-2D55E2, Treasure Case 2012 T722)each contain a pair of copper-alloy mould valves for palstaves, dating to the Taunton phase (c.1400-1275 BC).The complexity and volume of Late Bronze Age hoards is once again made apparent in Figure11.One pattern which stands out is the dense concentration of axe hoards across south-east Walesparticularly over the Vale of Glamorgan and Monmouthshireextending also into Herefordshire.In south-east Wales these axe only and axe dominant hoards account for 76% (n=32 of 42) of Late Bronze Age hoards, all of which are dated to the Ewart Park phase.Metallurgical hoards are particularly numerous across south-eastern England, accounting for 43% (n=36 of 82) of all Late Bronze Age hoards from Kent, Essex, and Suffolk.Outside of these core areas, the distribution of metallurgical hoards along the southern coast of south-western England and west Wales is also significant, with all 15 containing copper/copperalloy ingots.In contrast, they are relatively rare from northern England and completely absent from the West Midlands, Scotland, and Northern Ireland.The largest of these northerly metallurgical hoards is the Ewart Park phase Driffield II hoard from the East

TABLE 1 .
SUMMARY OF THE NUMBER OF BRONZE AGE HOARDS DISCOVERED PRIOR TO 1997 AND IN 1997-2021, BY CASE STUDY REGION AND

TABLE 2 .
SUMMARY OF BRONZE AGE HOARDS REPORTED IN 1997-2021 Roberts et al. (2013)Needham et al. (1997);Roberts et al. (2013)C.J. Griffiths.CHRONOLOGY, DISTRIBUTION & COMPOSITION, RECENT BRONZE AGE HOARDS, BRITAIN & N. IRELAND those Bronze Age hoards included within this dataset by sub-period of the Bronze Age (ie, Early, Middle, and Late).Discoveries of Early Bronze Age hoards are relatively rare and have a largely western and northern distribution.The Westenhanger hoard, Kent (PAS KENT-0330CE; Treasure Case 2019 T962), is the only example of an Early Bronze Age hoard within this dataset which has been discovered from eastern Britain but it is important to note that it actually fits well within the expected distribution of Arreton phase hoards which are concentrated in south-east England(Needham 2006, fig.38).The vast majority of hoards are, however, dated to the Middle and the Late Bronze Ages, whose distributions are also depicted in Figure 5. Middle Bronze Age hoards are overwhelmingly concentrated in southern England, particularly across Wiltshire, Hampshire, the Isle of Wight, Kent, and East Anglia.By comparison, Late Bronze Age hoards are far more numerous and widespread in their general distribution but with concentrations over parts of Kent, Essex, Suffolk, Norfolk, and south-east Wales also extending into southern Herefordshire and southern Powys.Smaller concentrations of Late Bronze Age hoards are also visible across other areas of England and Wales, including the south-western tip of Cornwall, the borderland of Shropshire and northeast Wales, and on land adjacent to Morecambe Bay in north-west England.The relatively few Late Bronze Age hoards across Gloucestershire and Wiltshireareas where Middle Bronze Age hoards have been frequently discovered and reported fromalso stand out when comparing the two maps in Figure

TABLE 3 .
SUMMARY OF THE VARIOUS HOARD CATEGORIES USED WITHIN THIS STUDY, ALONG WITH THE OBJECT TYPES WHICH CONTRIBUTE