Produce, Repair, Reuse, Adapt, and Recycle: the Multiple Biographies of a Roman Barrel

European Journal of Archaeology 23 (3) 2020, 356–380

Produce, Repair, Reuse, Adapt, and Recycle: The Multiple Biographies of a Roman Barrel

1 2 ROB SANDS AND ELISE MARLIÈRE 1School of Archaeology, University College Dublin, Ireland 2Antiquarium-Ibiza, Spain, and UMR 7041, University of Paris X, France

By the time the Roman empire reached its greatest extent, in the early decades of the second century AD, wooden barrels were a key part of a trade network that supported a complex extended economy. These objects do not, however, routinely survive in the archaeological record and very few sites have yielded large, multi-phase, assemblages for study. Although relatively rare, individual finds and assemblages have been found sufficiently regularly to allow us to consider barrel production and use during the Roman period. These objects can have complex cultural biographies from their original production to their final deposition. Current and previous research at Vindolanda, a Roman fort in northern Britain at the edge of the Roman empire, provides a context for reflection on these objects and their biographies. Emphasis is given to whether this material demonstrates repeated, possibly habituated, practices of adaption and recycling.

Keywords: Roman period, cooperage, Abies alba, Picea abies, Larix decidua, object biography, recycling

INTRODUCTION (e.g. Wilson, 2009; Rubio-Campillo et al., 2017). The term barrel has a more recent One aspect of the rhythm of life in a technical association with a certain size Roman fort, and indeed at various loca- of container, but, for the sake of con- tions across the empire, was the regular venience, the words cask and barrel are arrival of goods and supplies. used interchangeably here. They are Commodities came in a variety of contain- objects designed to enable movement, ers, of which amphorae, being made of their bellied form facilitating the rolling more durable material, are most frequently and rotation of a potentially large and recovered from archaeological contexts and heavy object. Studies of barrels in later consequently well studied and extensively periods show that even before arriving at published. However, barrels, i.e. liquid their first destinations casks may already tight wooden containers (Figure 1), also have had rich, complex lives (e.g. Ross, played a key infrastructural role in the 1985), and once in circulation could have Roman world (e.g. Bevan, 2014), and sig- multiple owners and pass through the nificantly contributed to a trade and hands of many people. supply network that extended across an Although organic objects survive less empire that reached its greatest extent by commonly in the archaeological record, the early decades of the second century AD evidence has been found sufficiently

Copyright © European Association of Archaeologists 2020 doi:10.1017/eaa.2020.7 Manuscript received 11 January 2019, accepted 6 February 2020, revised 14 June 2019 Downloaded from https://www.cambridge.org/core. University of Athens, on 30 Sep 2021 at 09:48:16, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2020.7 Sands and Marlière – Produce, Repair, Reuse, Adapt, and Recycle 357

Figure 1. Illustration of the terms used for the various parts of a cask. Note position and presence of some elements can vary.

Downloaded from https://www.cambridge.org/core. University of Athens, on 30 Sep 2021 at 09:48:16, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2020.7 358 European Journal of Archaeology 23 (3) 2020

regularly in Roman contexts (Figure 2) for biography can work alongside approaches barrels to be studied as a class, initially by such as affordance, material engagement, Ulbert (1959) and more recently by and entanglement to explore casks as Marlière, who drew all the strands of evi- objects, the people that made them, and dence together in an extensive synthesis the individuals that used and reused them. (2002), a study expanded by other Admittedly, this can be critiqued as a researchers in the last few years (e.g. Frei- somewhat ‘pick and mix’ strategy, but, to Stolba, 2017). Studies have tended to put it more formally, it adopts the brico- focus on cataloguing the evidence, devel- lage approach discussed by Olsen (Olsen, oping typologies of the original cask 2010:12–14) and accepts Cooney’s prag- forms, and considering origins, production matic observation that these approaches and primary use. This article builds on occupy a spectrum of thought with ‘sig- this work to explore the potentially nificant shared theoretical ground’ (2016: complex extended cultural biographies of 16 online first). these objects. Data from the northern frontier fort Since the publication of Kopytoff’sin- and settlement of Vindolanda provides the fluential essay on the biography of things case study considered here. Located in (1986), it has become standard to speak in northern England, 1.4 km south of terms of an object’s biography and to use Hadrian’s Wall, approximately half way this to reflect upon how an object’smeaning between Carlisle to the west and may change over time (e.g. Gosden & Newcastle to the east (Figure 2), it was Marshall, 1999;Joy,2009). Here we occupied almost continuously, from c. AD attempt to model some of the complexity 85 to c. AD 400, by various auxiliary units inherent in the biography of a cask at different stages of its existence (Birley, (Figure 3). We pay special attention to the 2002, 2009; Birley et al., 2016)(Table 1). later stages, examining how reuse, adaption, The site conditions have allowed for the and recycling are observed in the archaeo- survival of relevant material across several logicalrecordandquestiontowhatextent phases and represents one of a relatively these might reflect normative practices. small number of sites that have produced In working through a biographical larger assemblages of wooden objects from model, a simple birth, life, and death nar- this period. It lies some considerable dis- rative begins to break down and is usefully tance from the nearest likely primary pro- augmented by other approaches (Hahn & duction sites and the native distribution of Weiss, 2013: 4; Joyce & Gillespie, 2015: the trees from which casks of this period 10–11). Trends in thinking about things were primarily made (Figure 2). This have highlighted a ‘turn to the material’, therefore provides a contrast to sites such which has been increasingly emphasized as Vindonissa (Fellmann, 2009) and the over the last twenty years (Hicks & vicus of Tasgetium (Benguerel et al., Beaudry, 2010:2–3). While this ‘turn’ 2012), on the Rhine in Switzerland, and cannot be simply characterized, it can be the Colonia Claudia Ara Agrippinensium seen to bring together a valuable ‘family of (Cologne in Germany), which, like ideas’ (Hodder & Lucas, 2017: 121). In Vindolanda, yielded substantial wooden particular, it centres around a key unifying object assemblages from the first and theme that posits that ‘people make and second centuries AD. Here we also explore use things and things make people’ if other items in the assemblage might be (Cooney, 2016: 16). Thus, a collection of the product of recycling of cask staves. In ways of thinking can be considered; doing so, we shall raise the question of

Figure 2. Sites mentioned in text and natural distribution of key species. Grey dots represent other cask finds. Background cask distribution from Marlière (2002); natural distribution of tree species from the European Forest Genetic Resources Programme (http://www.euforgen.org/); major rivers from the European Environment Agency European catchments and Rivers network system (Ecrins); extent of the Roman empire in AD 117 from the Ancient World Mapping Center (http://awmc.unc.edu). All data sources freely available for use with acknowledgement.

Figure 3. Modelling the complex biography or itinerary of a barrel.

whetherwecangobeyondaconsideration context and considering the original form of material affordance, as first articulated by of the object, the material, and the pro- Gibson (1979), to more socially mediated duction process. Stave-built containers are approaches (e.g. Knappett, 2005). The known from earlier periods (e.g. Earwood, material from which a cask is made certainly 1993: 169), but these were open-topped affords a range of physical possibilities, but vessels, referred to in more recent times as did social habit and practice result in a con- white cooperage (Kilby, 1971:43–46). sistent range of possibilities? Techniques for creating closed casks can be divided into dry cooperage (Kilby, PRODUCTION,MATERIAL, AND USE 1971:46–53) for dry or semi-dry goods, and wet cooperage for liquids (Kilby, Before considering the latter stages of a 1971:53–64). Dry cooperage is less exact- cask’s life, it is worth putting this in ing than wet cooperage and, for certain

Table 1. Phases of occupation at Vindolanda (Vindolanda Trust). Period Date Range Unit

O AD 79–85 Possible occupation by unknown unit

I AD 85–90 Coh I Tungrorum

II AD 90–100 Coh I Tungrorum -> Coh VIIII Batavorum

III AD 100–105 Coh VIIII Batavorum

IV AD 105–120 Coh I Tungrorum

V AD 120–130 Coh I Tungrorum

VI AD 130–165 Possibly Coh II Nerviorum

VIA AD 165–200 Unknown

VIB AD 200–212 Unknown

VII AD 213–300 Coh IV Gallorum

VIII AD 300–367 Coh IV Gallorum

VIIIA AD 367–408 Unknown

IXA AD 409–500 Unknown

IXB AD 500–600 Brigomalos War band

X AD 600+ Unknown Christian Community

functions, results in less robust casks, Rhine-Westphalia, Germany; Hopf, 1967) sometimes only destined for single use and resinous material found on the surface (Edlin, 1973: 49). All three types of coop- of staves from Harelbeke in Belgium erage are present in Roman contexts with (Viérin & Léva, 1961) have often been substantial evidence for wet cooperage. cited as proof of wine but this may not be This technique is considered to have origi- definitive (Benguerel et al., 2012:55–60). nated in Gaul, although this is still Whether this was the norm has been debated (Desbat, 1997: 117–18; Marlière, questioned (Sealey, 2009: 25), and it has 2014: 48). Based on current indicators, been suggested that production centres in such as tree species, finds of tools, and the Gaul initially produced casks for ale, for representation of cooperage on funerary which there is increasing archaeological monuments, several large cooperage evidence, their use being extended to wine centres have been proposed in Gaul in the imperial period (Marlière, 2001: (Marlière, 2002: 177). Irrespective of the 186). Casks were produced in a variety of origin of the craft, there was a requirement sizes (see below) and were therefore likely to to produce containers of this type at a contain a variety of commodities. For potentially large scale and such production example, the casks may have contained was highly developed and extensive by the foodstuff including fish (Sciallano, 1993: 18; time a northern frontier fort like Marlière, 2002: 174), or olives, or olive oil Vindolanda was first occupied. (Benguerel et al., 2012: 269). Irrespective of Classical authors, such as Strabo and the specific identification of contents, the Pliny, indicate that wine was stored or development of specialized bulk containers transported in barrels; moreover, the pres- such as these is argued to have been driven ence of bung and vent holes on many by the need to move refined products such examples indicate that liquids were present as wine and oil (Bevan, 2014: 388). (Marlière, 2002: 173). Analytical evidence In more recent times, wet cooperage of tartrate residues from Oberaden (North required long apprenticeships of up to

seven years or longer to achieve master organizations formally underlined the status (see Kilby, 1971). Such training social identity of individuals who identi- imparted and maintained the highly fied themselves as ‘coopers’, and who embodied skills necessary to ensure con- might in turn be identified by others as sistent, quick, and reliable production. such. In a first- or second-century AD While there is as clear chaîne opératoire context, the expression of a craftworker’s with a series of essential steps (Marlière, identity through social structures is 2002:29–40), which may perhaps be rec- evident in the presence of collegia, ognizable regardless of cultural or chrono- although the exact role these played con- logical context, the exact way these steps tinues to be debated (Verboven, 2011), were followed is arguably more a product and so far there is no evidence of a colle- of learnt habit and practice, driven by cul- gium of coopers. That a cooper had a tural context. Habituated modes of prac- recognized identity within the Roman tice, learnt from existing practitioners, world is, however, demonstrated by at sustained a consistency of form and helped least half a dozen funerary monuments, to maintain the cultural identity of the including an example from Berbourg in practitioner. While some simple measur- Luxembourg that evokes the interior of a ing instruments, such as a pair of com- cooperage workshop, displaying tools on passes, are used to assess the dimensions the wall, barrels, and workers (Marlière, of the head, much of the work is gauged 2002: 120–21). ‘by eye’, implying increasingly heightened The species of tree chosen by a cooper levels of embodied cognition with growing must afford the necessary balance of phys- experience (see Malafouris, 2013:57–58, ical properties required to produce a and 209 for a consideration of material useable stave-built vessel for a given culture and the extended mind; purpose, this restricts what timber might Marchand, 2010 for a discussion of non- be most suitable in each situation. verbal skill transfer and aquisition amongst Nevertheless more than one species could fine woodworkers). In this sense, the fit the necessary basic criteria and conse- materials, the form of the object, the tools quently specific traditions of production used, and the specific processes involved could emerge based on both practical and are potentially very tightly bound with cultural choice. It is arguable that it can identity; they are ‘entangled’ in the fullest become a cultural norm to associate an sense. The sum of humans depending on object, such as a barrel, with a specific things, things depending on other things, material. An emerging preference could be things depending on humans, and humans reinforced and maintained through the depending on humans (Hodder, 2014: passing on of craft skills across genera- 19–20) can be argued to be particularly tions, creating an increasing acceptance of apparent in highly specialized crafts such ‘the correct way’ for a cask to be made. as this. The choice may be influenced by other From at least the medieval period in factors such as the aesthetic, symbolic, or Europe, such crafts became more defined other properties of the timber chosen. The and prescribed through formal organiza- chemical properties of the wood and pro- tionssuchasguilds(e.g.Foster,1944; duction processes, for example, can impart Mair, 2013), which further reinforced additional qualities, such as flavour vari- and maintained the understood, ‘correct’ ation, to some types of stored contents way of conducting the craft, through such as wine (Ribereau-Gayon, 1994). In rules, rituals, and rites of passage. Such a potentially complex feedback loop, the

desire for, or social acceptance of, aspects the explanation given is primarily asso- such as certain flavour notes might both ciated with the use of silver fir in larger further reinforce and mediate material construction projects. It is proposed here, choice. Modern wet cooperage in Europe, however, that barrel production must also for example, predominantly uses oak have had a significant impact. Modern (Ribereau-Gayon, 1994: 8), which has estimates suggest that an oak tree around specific effects on aspects such as the 200 years old and with a diameter of c.90 colour and flavour of a stored liquid, such cm at chest height can, on average, as wine. While oak was used for smaller produce two complete casks of standard coopered items in the early centuries AD, size containing some 200–300 litres and may have begun to be used more (Edlin, 1973: 98; Marlière, 2002: 29). It is regularly in the later centuries of the worth noting that surviving Roman barrel Roman empire (e.g. Rule & Monaghan, capacity could be much greater than this 1993), it was not the material normally modern standard (Marlière, 2002: 165). associated with barrel production. Additionally, not every tree will be of a Through much of this period, coopered sufficient quality; in a recent article analys- casks were predominantly produced from ing yields suitable for cooperage in Spain, fir or spruce/larch, with a possible shift only ten per cent of the logged oak was from just fir to a mix of species over time useable (Riesco Muñoz et al., 2013: 696). (Marlière, 2014: 53). This choice also con- The need for quality timber and the fact trasts with, for example, Iron Age and that wet cooperage required precise skills Roman Britain (e.g. Horn, 2015; Sands & meant that the casks produced were of Horn, 2017), early medieval, Viking and enough value to warrant extending their medieval Ireland (e.g. Comey, 2003, life through repair. Modern estimates put 2010), and early medieval Britain (e.g. a barrel’s longevity at around eight years, Cook, 2004; Comey, 2013) which all have during which time repair and possibly dis- strong traditions of using yew for coopered assembly and readjustment might be items, with possible aesthetic and ritual expected (Marlière, 2002: 37). There is motivation. While most of these are open- evidence for such activity in, for example, topped vessels, some were clearly designed the Roman town of Calleva (modern to hold liquids and there is rare evidence Silchester; Boon, 1974: 264) and it would for yew cask staves from early medieval have required a defined skill set. In more sites in Ireland (e.g. Linnane & Kinsella, recent times, coopers could be employed 2009: 115), although whether these primarily for this purpose. The presence of represent wet or dry cooperage has yet to a cooper in London in AD 80–90/95 indi- be determined. cates that local skills were needed, and In recently examined assemblages of the available, for a range of activities that Roman period, silver fir not only tends to probably included repair, reassembly, and predominate relative to spruce/larch but adaption (Tomlin, 2016: 86, writing can also make up a significant proportion tablet 14). of the total species range recovered, both coopered and non-coopered (Figure 4). The exploitation of this resource has been REUSE,REPURPOSING,ADAPTION, AND suggested as the primary cause of its RECYCLING decline in pollen and charcoal records during the first few centuries AD (e.g. Archaeologically, reuse and recycling have Küster, 1994; Nakagawa et al., 2000) and been repeatedly considered by Schiffer,

Figure 4. Proportions of fir, spruce/larch, and undifferentiated softwood species in three of the larger wooden assemblages of the period. Outer circle: Vindolanda; middle circle: Cologne (Tegtmeier, 2016); inner circle: Vindonissa (Fellman, 2009).

first in his examination of archaeological considering how the recycling of specific context (1972) and latterly in his materials is analytically recognized and Behavioral Archaeology(2010). Four basic quantified, such as in Freestone’s examin- behavioural processes that fall under the ation of Roman glass (2015), or quantify- broad term ‘reuse’ are suggested: lateral ing metal flow and recycling of copper- cycling, secondary use, recycling, and based metals in Britain in the first millen- conservatory processes (Schiffer, 2010: nium AD (Pollard et al., 2015), to eco- 32–33). These encompass ways in which nomic considerations such as Fleming’s an object might move from one person to assessment of recycling in Britain after the the next, retaining its original use, keep fall of Rome’s metal economy (2012), or the same form but change use, be altered extended biographies of individual artefact to have another use, or be curated. types, such as Swift’s examination of rings Academic discussion of these themes has made from Romano-British bracelets had increasing traction in recent years, in (2012) or Abdelhamid’s work on the part, no doubt, because of the raised extended lives of amphorae (2013). profile of recycling in broader modern dis- Inevitably, the focus has predominantly course. The discussion has ranged from been on inorganic materials such as glass,

metal, and pottery. Wood has received Repurposing is used when the primary relatively limited attention in this context, purpose of the cask has shifted but the most commonly in terms of observing form is retained. This could simply be a reuse, particularly in structures (e.g. Ayre change in what it contains, such as in the & Wroe-Brown, 2015), and occasionally case of the Grado wreck at Gorizia in in relation to wooden objects, of which Italy, which sank at the end of the second Goodburn’s study of material in Roman century or beginning of the third century London is most relevant here (see below). AD, containing a cask full of broken glass, While Schiffer’s lateral cycling could presumably destined for recycling apply to all objects, depending on social (Auriemma, 2000: 27). Alternatively, norms, the exact nature of other responses repurposing could be more dramatic, such depends in part on the possibilities as barrels used in the lining of well shafts. afforded by the different materials. Thus, Adaption refers to a coopered structure while all objects could be subject to lateral being retained but the form physically cycling or secondary use, the material altered, such as cutting down a barrel to attributes may allow or constrain certain produce a tub. Recycling is reserved for responses. At a basic level, glass is fragile situations in which the material from and shatters but can be melted down; individual components has been used to metal cuts, bends, breaks, and melts; produce other objects. pottery breaks and can be reshaped to a limited extent. Wood, on the other hand, burns, snaps but does not shatter, and VINDOLANDA – ANORTHERN FRONTIER cannot be melted and reformed. It can, CASE STUDY however, be carved, sawn, split, shaved, and bent, attributes that are mediated by Amphorae, along with other less durable the differing anatomies of different objects such as skin containers, sacks, and species, which make it possible to create a casks, would have been part of the con- potentially wide range of new forms. temporary rhythm of life. They provided a Furthermore, in the case of a barrel, as a visual signature of the coming and going composite object, it can be taken apart, of materials and goods, connecting a fort rebuilt, and adapted. like Vindolanda to other settlements, to Schiffer’s scheme continues to be a nearby centres, and to the wider empire. useful basis, but it is inevitably an Both sacks and barrels are mentioned in approximation of the complexity of actual the contemporary correspondence at practice (Schiffer, 2010: 32). For the pur- Vindolanda (e.g. Tab. Vindol. II 180, poses of this article, the verbs reuse, Bowman & Thomas, 1994) and the site repurpose, adapt, and recycle are used to has sealed anaerobic conditions suitable capture distinct aspects of the potential for the survival of organic material. Such biographies of a cask (Figure 3). Reuse deposits occur from Period I to Period here is reserved for a situation in which a VII, with the bulk of the material surviv- cask is reused for essentially its original ing in periods II to V, largely encompass- purpose in the broadest sense (e.g. to ing the reigns of the emperors Trajan and contain a liquid); evidence from multiple Hadrian. To date, direct evidence for sites indicates that barrels could be filled casks comes from 390 surviving coopered and refilled, with overlapping branding elements, representing the dominant arte- marks providing proof of reuse, move- fact class and occurring in all phases that ment, and possibly changing ownership. have the appropriate preservation

conditions. This assemblage is being (Lodwick, 2017:14–15), it is highly examined within a larger project analysing unlikely to have produced the size or portable wooden objects from the site quantity of material required for the sus- (Sands & Hather, in prep.) and comple- tained production of casks. ments initial studies of a sample of fifty It is worth noting that in more recent staves examined by Reis (1997) and a tar- times the timber trade has treated silver fir geted assessment of amphorae and cask and Norway spruce as one and the same, types by Marlière (2003; Marlière & referring to both as ‘white deal’ (Edlin, Costa, 2005). 1973: 129), and Roman finds include evidence that the two species were used in the same cask, as in examples from Aardenburg A question of material and Harelbeke in Belgium (Frison, 1961) and Rijswijk in the Netherlands (Casparie, Archaeological wood samples can be iden- 1978). tified to genus or species based on differences in anatomical structure as defined in standard reference works (e.g. IAWA Vindolanda, an edge of empire Commitee, 1989; Schweingruber, 1990; destination Hather, 2000). Using this method, we have so far identified 276 items made Few, if any, casks reached as far as a nor- from fir, which, given its European thern frontier fort without having passed context, is narrowed down to silver fir through several hands, including the trans- (Abies alba Mill.) (Mauri et al., 2016). In fer of ownership, probably having been addition, forty-two items have been iden- emptied and refilled several times, at times tified as made from either spruce or larch, with different contents. Multiple and which in our context is probably either overlapping branding marks on some Norway spruce (Picea abies Karst.) staves provide strong physical proof of a (Caudullo et al., 2016) or European larch complex biography (Figure 5), and similar (Larix decidua Mill.) (Da Ronch et al., branding is also common on other sites 2016). Spruce and larch are considered by (e.g. Etter et al., 1991: 110). Barrels could some to be anatomically indistinguishable, also be deliberately taken apart, repaired, especially in archaeological wood (e.g. and rebuilt. Some examples at Vindolanda Chabal & Feugère, 2005: 182). have a second croze groove, probably Differences have, however, been suggested showing either repair prior to arrival or and features observed by Batholin are con- further reuse while at the fort. sidered the most reliable (Batholin, 1979; In addition to branding, marks can be Anagnost et al., 1994). Work at scored into the barrel surfaces and there is Vindolanda is beginning to indicate that a suggestion, just as with terra sigillata spruce rather than larch is present, but stamps, that the original makers were also further research is required. identifying themselves on the casks. At None of these species, fir, spruce, or Tasgetium, for example, the verb fecit (‘he/ larch, has a natural distribution that she made’), incised in cursive script, is includes Britain (Figure 2) and the nearest found on some barrel heads and it is pos- natural occurrence is some 1000 km south sible that the same workshop was also of Vindolanda. Notwithstanding that a responsible for casks found at Vitudurum strong tradition of Roman silviculture (Benguerel et al., 2012: 260). Production allowed for the growth of spruce in Britain signatures have not yet been definitively

Figure 5. Tub stave (W-1992-1098) with overlapping brands and original barrel bung in situ.

identified on the Vindolanda material, but of correspondence from Vindolanda that cursive script is present on some of the describes the redistribution of wheat. The cant and middle staves. tablet starts with the following words: Marlière suggests five size classes of ‘Account of wheat measured out from cask across the empire (2001: fig. 103; that which I myself have put into the 2002: 159–67) and identifies three of barrel’ (Tab. Vindol. II 180; Bowman & these amongst the Vindolanda material: Thomas, 1994). The wheat accounted for is 320 modii, with a modius equiva- . Group 1: tonnelets – small casks, c. 30– lent to approximately nine litres dry 50 cm high measure, a figure suggesting that, even . Group 2: barriques – casks, c. 100 cm accounting for some very large Roman high barrels, more than one would have been . Group 5: fûts – extremely large casks, c. required. That barrels were probably 200 cm high. reusedasstoragecontainersisalsosug- That very large casks were present is gested at other sites, such as Vindonissa, suggested by the survival of one complete where barrels, recessed into the floor of a middle stave from a barrel head, measur- possible store house, have provided evi- ing 1.34 m. Repurposing of larger barrels dence for pomegranates (Bakels & is indicated by one of the surviving pieces Jacomet, 2003:553).

None of the casks in the Vindolanda assemblage survive in their complete articulated state, although some individual elements retain their original form, especially the cant and middle staves from heads or less frequently complete side staves. There is, by contrast, substantial evidence that many casks were adapted or recycled, with stave offcuts and staves from cut-down open tubs predominating the assemblage.

Offcuts Figure 6. Offcut (W-1988-573) sawn from a cask stave. One clear sign that casks were being reused or adapted is the waste from those processes, and there are eight examples be cut and staves refitted to a base. from Vindolanda (Figure 6). Although However, complete barrels have been this is a small sample, such offcuts are recovered with bung holes closer to the more likely to be discarded and may never barrel heads (e.g. Kühlborn, 1992: 100, enter the archaeological record, having, for fig. 37) and thus the Vindolanda practice example been destroyed through deliberate might more routinely have been cutting burning. The examples that have survived casks in half in the case of larger tubs or are found in multiple phases (I–IV), which cutting the top and the bottom, leaving adds weight to the argument that adaption waste material from the centre in the case of casks was a relatively routine activity. of the shorter tubs. Even in this simpler scenario the individual staves were being reshaped to provide a new internal Adaption of casks to create other chamfer, an attribute also noted on a simi- coopered forms larly adapted tub stave from the fort of Segontium in North Wales (Boon, 1975: It is argued here that Vindolanda unam- 54, fig. 2), and some skill would have biguously demonstrates a routine adaption been required to create a useable tub. of the larger casks (Marlière’s Group 5) to Making tubs is not just a unique produce open-topped tubs. That these response to a chronologically specific derived originally from casks is shown not requirement. The phases from which the only by the thickness and general form of evidence derives clearly shows that such the staves but also by the presence of activity was repeated and could be partial or complete bung holes in some of regarded as habituated, at least across this the shorter tub staves (Figure 6). Many period at this location (Figure 7). No barrel finds have these holes close to the period, where significant organic preserva- centre of their girth (e.g. Boon, 1975: pl. tion was present, privileged this behaviour VII), which suggests that more complex more than another and this activity was conversions might have been undertaken taking place irrespective of the specific at Vindolanda. If this is the case, the auxiliary unit occupying the fort. adaption may imply some local coopering Furthermore, examination of the dimen- skills, as new croze groves would have to sions of the cut-down items shows

Figure 7. Coopered elements by Vindolanda phase. A: all coopered items; B: only head staves, complete cask staves, and bungs; C: only adapted tub staves and offcuts. Normalized as a proportion of wooden finds in each phase.

consistency in the way this type of conver- and IV. The Segontium stave, noted sion was carried out (Figure 8), with the above, is 28 cm in height, well within the majority falling within a group of shorter predominant range of the shorter tubs items between 24 and 40 cm high, a less found at Vindolanda. A similar style of frequent middle group between 48 and 72 object is also known at the vicus of cm high, and a few tall tubs around 90 cm Tasgetium, where a small tub in a partially high. There is a spike in the distribution articulated state has an average stave at around 32 cm, and it is worth noting height of 21.5 cm (Benguerel et al., 2012). that this is not explained by multiple Exactly why these tubs were required is staves from the same tub as there is almost not clear; however, as this is a repeated, equal representation from phases II, III, arguably standardized practice, it suggests

Figure 8. Frequency distribution showing the height of tub staves at Vindolanda.

a recurrent need. There would have been, be argued to have marks characteristic of for example, a consistent need for leather. horses chewing a wooden container in the Analysis of adapted barrels at the vicus of expectation that it would be filled. On a Vitudurum (Oberwinterthur, Winterthur fort where cavalry would have been an in the canton of Zürich, Switzerland) pro- important part of daily life such uses vided evidence for traces of animal hair, might also be expected; significantly, bristles, and leather or skin residues, sug- perhaps, Segontium, where there is evi- gesting that they were employed in dence for similar tubs, also housed an tanning (Clerici, 1983: 21). A similar sug- auxiliary cavalry unit. gestion has been made for some of the Although evidence for the adaption of sixteen ‘cask pits’ found at Tasgetium the larger barrels predominates, the adap- (Benguerel et al., 2012: 268). However, tion of smaller barrels is also documented. definitiveevidenceofthepresenceoftan- Open-topped containers, such as buckets, neries has proved elusive (Driel-Murray, and smaller (approximately 20 cm high) 2011) and, while tubs of such dimensions straight-sided or slightly tapering casks are could have been used in this process, the both considered to have been produced finds from Vindolanda cannot be defini- from staves originally derived from smaller tively assigned to it. This may not, cask types. The latter suggest the local however, be the only explanation and, presence of at least some dry coopering significantly, some of the stave ends can skills (Figure 9).

Other objects of fir, spruce, or larch: either fir, spruce, or larch (e.g. Pearce, imports, opportunistic, or routine 2004: 45; Saedlou & Dupéron, 2007; recycling? Benguerel et al., 2012; Hartmann, 2015; Stewart, 2016; Tegtmeier, 2016:21–77). Casks represent a ready-to-hand source of The choice of these species has merit, a raw material, one that was, in this case, since a recess can be easily split out and not available locally, and that had certain the surface left would have had a rougher material affordances. The barrel staves texture facilitating the keying of the wax. potentially provided a seasoned, pre-cut, At locations away from the natural distri- predominantly radially converted source of butions of these species, the ‘exotic’ nature softwood timber, of relatively uniform and of the wood has been argued to show that predictable dimensions. It would not be the stylus tablets were manufactured else- surprising to encounter some use of this as where and reached their destination ready- a source material, especially when it is not made (e.g. Bowman, 1994: 47; Austin, available locally. Items made from the 2015: 16). Saedlou and Dupéron (2007: same tree species as a barrel currently 85), suggest that this implies either a include drop lids, wax tablets (tabulae large-scale trade in this type of wood, with ceratae), framing pieces, and domestic local production or centres of production utensils. near the necessary raw material, the fin- Drop lids are a simple form of square, ished article then being distributed further unhinged lid that has a thick central afield. While the presence of tablet pro- section with a thin lip. Four of these have duction centres is likely, especially as a been identified at Vindolanda, all made large and consistent supply would have from fir or spruce/larch. All four are rela- been required, this does not preclude local tively roughly made, matching closely the manufacture if a source of material was thickness of staves from the larger barrels available. The site of Bloomberg in and, in at least one example, there is a London provides a valuable touchstone suggestion of the original lower chamfer, here. As at Vindolanda, written texts with one edge created by utilizing the cut survive and include a number of tabulae of the croze groove (Figure 10a). These ceratae, although the date range at objects, although relatively rare, have so far Bloomberg is earlier (50s to 90s AD). been identified across three phases (II, III Significantly, Goodburn has argued that and IV). A drop lid, identified as made barrel parts were used to produce stylus from spruce, is also known from Roman tablets, using the presence of offcuts from contexts at Castle Street, Carlisle (Padley, the ends of staves and wood chips as 1991: fig. 183). The choice of species is proxy evidence for this process not functionally essential for this simple (Goodburn, 2016:fig.7).Heconjectures form, as a later example in oak from that, with the ends removed, the central Hiberno-Norse York demonstrates part of the stave could be cut into tablet- (Morris, 2000: 2292). sized blocks and subsequently split into A small selection of stylus tablets from individual blanks, estimating that from Vindolanda has previously been micro- one large cask in good condition as many scopically examined and identified as as 420 tablets could potentially be manu- spruce/larch (Bowman & Thomas, 1983: factured (Goodburn, 2016:9).Directevi- 26–29). This fits with other finds from dence for tablet production is very rare across Europe and Britain, most of which and a find from Cologne indicates that have also been identified as produced from blanks could be produced by sawing

Figure 9. Adapted small barrel staves (W-1992-1069).

rather than splitting (Tegtmeier, 2016: indicating repeated use. Consequently, even fig. 35). if casks were being reused in local produc- Goodburn’s hypothesis is attractive, and tion, either the process was sufficiently one that had already been positively consid- labour-intensive or the second-hand mater- ered for the Vindolanda material. As a ial not consistently available such that thought experiment, at the very least, it stylus tablets were worth curating. counters an automatic assumption that Three items (W-1986-77, W-1993- tabulae ceratae, or indeed other objects 1200, and W-1993-1263) are parts of made from spruce/larch or fir, must be pro- frames and all are made of the same soft- duced at a distance or require movement of wood species as the surviving coopered the raw material. It should however be elements (Figure 10b). All three items are noted that there is no absolute proof that from period III and it is possible that they this was part of the recycling behaviours at represent window frames from the com- Vindolanda: while there are stave offcuts manding officer’s residence. More recent present, like those found at Bloomberg, excavations have uncovered a fourth there is no direct connection to the produc- example (W-2005-06A) from a context tion of tabulae ceratae other than that they that has only a broad date range (c. AD are produced from the same species of tree. 92–140); it is therefore not impossible that Furthermore, at Vindolanda there is evi- it was part of the same higher-status dence for extensive reuse of the tablets, building as the first three frames. While with multiple texts superimposed, similar objects have been found in Britain

Figure 10. a. Drop lid (W-1988-607), b. Framing element (W-1986-77), c. Spatula (W-1991-732).

(e.g. Blurton, 1977: 67, fig. 20), the evidence at Vindolanda that these items largest group, from anywhere in the were made from recycled cask parts, and Roman empire, comes from Vindonissa hence this remains conjecture. (Fellmann, 2009:94–96, 173), where 163 Nonetheless, these items can no longer be items have been categorized as assumed to be either imported complete or ‘Fenstersprosse’ or window bars. There, all produced from imported raw timber. items were identified as manufactured What is perhaps more significant is that, from fir. Sixteen were also recovered from even on the northern frontier, the material Roman contexts in Cologne, of which two deemed to be appropriate for this specific were identified as spruce and the remain- purpose is a wood like silver fir, a timber der, apart from one unidentifiable piece, as only available locally from recycled objects fir (Tegtmeier, 2016: table 23). As with and used even though other types of wood the stylus tablets, there is no direct could serve the same purpose.

The final object made from silver fir at while this reading of the evidence has Vindolanda is a single spatula, which merit, it is argued here that some of the belongs to a broader range of domestic practices of recycling, reuse, and adaption utensils identified in the assemblage were also systematic and habituated. This (Figure 10c). It is a simple form, readily may be connected to the origins of the made from radially cleft timber, and at casks arriving at a fort like Vindolanda Vindolanda this artefact type is more after a long and complex journey and may usually made from locally available woods. partly echo the range of things that were The ease with which these objects could most usually produced from these particu- be made and their simple form suggest lar tree species closer to their natural that this one example may well have been origin. While an object and its material produced from wood recycled from casks. had certain affordances, the usual If this is the case, it is plausible that more responses were also defined, at least in opportunistic use also occurred. part, by what was deemed most appropriate to create from it, and this could be socially framed. A line in Kopytoff’s ori- CONCLUSION ginal discussion of the cultural biography of things encapsulates some of these inter- In this article we have presented a model acting influences: ‘Some of this clash that explores the cultural biography of between culture and individual is inevit- barrels, focusing on the Roman empire in able, at least at the cognitive level. The the first and second centuries AD, paying world of things lends itself to an endless special attention to the physical evidence number of classifications, rooted in natural for the latter stages of their existence. A features and cultural and idiosyncratic per- cask had an embodied value, albeit one ceptions’ (Kopytoff, 1986: 76). In our that could change over time; at a basic case, amid the social interactions, the level its manufacture required resource, rituals, the commercial and military energy, materials, training and continued imperatives, the reuse, adaption, repurpos- practice. This sense of ‘value’ promoted ing, and recycling of objects like wooden longer life uses, repair, and multiple reuse, casks would have been a necessary, inher- and this is clearly visible in the archaeo- ently understood part of the rhythm of logical record. The materials from which daily life. casks were made could also provide a rationale for adaption and recycling, and, in the case of a northern frontier fort, pro- ACKNOWLEDGEMENTS vided a source of material not naturally present. From this perspective, the beha- Thanks to the Vindolanda Trust for access viours observed at Vindolanda and else- to the material, especially to Robin Birley where could be read from a purely and Patricia Birley and, latterly, Andrew economic, or instrumentalist, viewpoint. Birley and Barbara Birley for their contin- The arrival of a cask at a fort or settlement ued support and insights. The team at such as Vindolanda would have provided Vindolanda continue to make our research opportunities for the reuse, adaption, and/ visits a genuine pleasure. We are grateful or recycling of a useful material that was for critical suggestions on drafts of this not locally available; the motivation for article by Anthony Birley, Andrew Birley, doing so might thus be framed as purely Barbara Birley, Neil Carlin, Lorna pragmatic and opportunistic. However, O’Donnell, Tadhg O’Keefe, Meriel

McClatchie, and the anonymous Bevan, A. 2014. Mediterranean Containerization. reviewers. Thanks also to Damian Current Anthropology, 55: 387–418. Goodburn and Karen Stuart, MOLA, for Birley, A. 2002. Garrison Life at Vindolanda: A Band of Brothers. Stroud: Tempus. valuable discussion. Finally, thanks to Jon Birley, A., Meyer, A. & Greene, E.M. 2016. Hather for collaboration and initiating the Recent Discoveries in the Fort and examination of the wooden small finds. Extramural Settlement at Vindolanda: Support of the UCD College of Social Excavations from 2009–2015. Britannia, – Sciences and Law Research Fund is grate- 47: 243 52. Birley, R. 2009. Vindolanda: A Roman Fort on fully acknowledged. Hadrian’sWall. Stroud: Amberley. Blurton, T.W. 1977. Excavations at Angel Court, Walbrook, 1974. Transactions of the REFERENCES London and Middlesex Archaeological Society, 28: 14–100. Abdelhamid, S. 2013. Against the Throw- Boon, G.C. 1974. Silchester: The Roman Town away-mentality: The Reuse of Amphoras of Calleva. Newton Abbot: David & in Ancient Maritime Transport. In: H.P. Charles. Hahn & H. Weiss, eds. Mobility, Boon, G.C. 1975. Segontium 50 Years On: A Meaning, and the Transformations of Roman Stave of Larchwood and Other Things. Oxford & Oakville (CT): Oxbow Unpublished Finds Mainly of Organic Books, pp. 91–106. Materials, Together with a Note on Late Anagnost, S.E., Meyer, R.W. & De Zeeuw, Barracks. Archaeologia Cambrensis, 124: C.D. 1994. Confirmation and Significance 52–67. of Bartholin’s Method for the Bowman, A.K. 1994. Life and Letters on the Identification of the Wood of Picea and Roman Frontier: Vindolanda and its People. Larix. International Association of Wood London: British Museum Press. Anatomists (IAWA) Journal, 15: 171–84. Bowman, A.K. & Thomas, J.D. 1983. Auriemma, R. 2000. Le anfore del relitto di Vindolanda: The Latin Writing-tablets. Grado e il loro contenuto. Mélanges de London: Society for the Promotion of l’école française de Rome, Antiquité, 112: Roman Studies. 27–51. Bowman, A.K & Thomas, J.D. 1994. The Austin, J. 2015. Letter Writing at Vindolanda Vindolanda Writing Tablets II (Tabulae (Northumberland/GB). In: M. Scholz & Vindolandenses II). London: British M. Horster, eds. Lesen und Schreiben in Museum Press. den römischen Provinzen. Mainz: Römisch- Casparie, W.A. 1978. Über die Holzarten der Germanisches Zentralmuseum, pp. 15–26. zwei römerzeitlichen Fässer von Rijswijk. Ayre, J. & Wroe-Brown, R. 2015. The In: J.H.F. Bloemers, ed. Rijswijk (Z.H.), Eleventh- and Twelfth-century Waterfront ‘De Bult’. Eine Siedlung der Cananefaten and Settlement at Queenhithe: Excavations (Nederlandse Oudheden, 8). Amersfoort: at Bull Wharf, City of London. Rijksdienst voor Oudheidkundig Archaeological Journal, 172: 195–272. Bodemonderzoek, pp. 438–46. Bakels, C. & Jacomet, S. 2003. Access to Caudullo, G., Tinner, W. & De Rigo, D. Luxury Foods in Central Europe during 2016. Picea abies in Europe: Distribution, the Roman Period: The Archaeobotanical Habitat, Usage, and Threats. In: J. San- Evidence. World Archaeology, 34: 542–57. Miguel-Ayanz, D. De Rigo, G. Caudullo, Batholin, T. 1979. The Picea-Larix Problem. T. Houston Durrant & A. Mauri, eds. International Association of Wood European Atlas of Forest Tree Species. Anatomists Bulletin,1:68–70. Luxembourg: European Commission, pp. Benguerel, S., Brem, H., Ebneter, I., Ferrer, 114–16. M., Hartmann, B., Leuzinger, U., et al. Chabal, L. & Feugère, M. 2005. Le mobilier 2012. Tasgetium II. Die römischen organique des puits antiques et autres con- Holzfunde (Archäologie im Thurgau 18). textes humides de Lattara. In: G. Piquès Frauenfeld: Department für Erziehung & R. Buxó, eds. Onze puits gallo-romains und Kultur des Kantons Thurgau. de Lattara (Ier s. AV. N. È. - IIe s. de N. È.).

Fouilles programmées 1986–2000. Lattes: Edlin, H.L. 1973. Woodland Crafts in Britain: Association pour la recherche An Account of the Traditional Uses of Trees archéologique en Languedoc oriental, and Timbers in the British Countryside. pp. 137–88. Newton Abbot: David & Charles. Clerici, R. 1983. Römische Holzfässer aus Etter, H.F., Brogli, R.F., Fellmann, R., Vitodurum. Helvetica Archaeologia, 53/14, Martin-Kilcher, S., Morel, P. & Rast, A. 14–24. 1991. Beiträge zum römischen Comey, M.G. 2003. Stave-Built Wooden Oberwinterthur. Vitudurum 5, Teil A: Die Vessels from Medieval Ireland. The Journal Funde aus Holz, Leder, Bein. Gewebe. of Irish Archaeology, 12/13: 33–77. Zürich: Orell Füssli. Comey, M.G. 2010. Coopers and Coopering in Fellmann, R. 2009. Römische Kleinfunde aus Viking Age Dublin (Medieval Dublin Holz aus dem Legionslager Vindonissa Excavations 1962–81, series B, vol. 10). (Veröffentlichung der Gesellschaft Pro Dublin: Royal Irish Academy. Vindonissa, 20). Brugg: Pro Vindonissa. Comey, M.G. 2013. The Wooden Drinking Fleming, R. 2012. Recycling in Britain after Vessels in the Sutton Hoo Assemblage: the Fall of Rome’s Metal Economy. Past Materials, Morphology, and Usage. In: M.D. & Present, 217: 3–45. J. Bintley & M.G. Shapland, eds. Trees and Foster, W. 1944. The Coopers’ Company: A Timber in the Anglo-Saxon World.Oxford: Short History of the Worshipful Company of Oxford University Press, pp. 107–121. Coopers of London. Cambridge: Cambridge Cook, J.M. 2004. Early Anglo-Saxon Buckets. University Press. Oxford: Oxford University School of Freestone, I.C. 2015. The Recycling and Archaeology. Reuse of Roman Glass: Analytical Cooney, G. 2016. Material Culture. In: A. Approaches. Journal of Glass Studies, 57: Gardner, M. Lake & U. Sommer, eds. 29–40. The Oxford Handbook of Archaeological Frei-Stolba, R. 2017. Holzfässer: Studien zu Theory. Oxford: Oxford University Press, den Holzfässern und ihren Inschriften im pp. 1–31. römischen Reich mit Neufunden und Da Ronch, F., Caudullo, G., Tinner, W. & Neulesungen der Fassinschriften aus De Rigo, D. 2016. Larix decidua and Oberwinterthur. Zürich: Baudirektion Other Larches in Europe: Distribution, Kanton Zürich, Amt für Habitat, Usage, and Threats. In: J. San- Raumentwicklung, Kantonsarchäologie. Miguel-Ayanz, D. De Rigo, G. Caudullo, Frison, E. 1961. Examen anatomique des bois T. Houston Durrant & A. Mauri, eds. du puits romain n˚ I de Harelbeke. European Atlas of Forest Tree Species. Latomus, 20: 800–05. Luxembourg: European Commission, pp Gibson, J.J. 1979. The Ecological Approach to 108–10. Visual Perception. Boston: Houghton Desbat, A. 1997. Le tonneau antique: ques- Mifflin. tions techniques et problème d’origine. Goodburn, D.M. 2016. The Manufacture of In: D. Garcia & D. Meeks, eds. Waxed Stylus Writing Tablets in Roman Techniques et économie antiques et rnedie- London. In: R.S.O. Tomlin, ed. Roman vales. Le temps de l’innovation. Paris: London’s First Voices: Writing Tablets from Errance, pp. 113–20. the Bloomberg Excavations, 2010–2014 Driel-Murray, C.V. 2011. Are We Missing (MOLA Monograph 72). London: Something? The Elusive Tanneries of the Museum of London Archaeology, Roman Period. In: R. Thomson, & Q. pp. 8–13. Mould, eds. Archaeological Leather Gosden, C. & Marshall, Y. 1999. The Group’s Conference ‘Have We Got a Cultural Biography of Objects. World Tannery? The Archaeology of the Skin Archaeology, 31: 169–78. Processing Industries’. London: Archetype Hahn, H.P. & Weiss, H. 2013. Introduction: Press, pp. 69–83. Biographies, Travels, and Itineraries of Earwood, C. 1993. Domestic Wooden Artefacts: Things. In: H.P. Hahn & H. Weiss, eds. In Britain and Ireland from the Neolithic to Mobility, Meaning, and the Transformations Viking Times. Exeter: University of Exeter of Things. Oxford & Oakville (CT): Press. Oxbow Books, pp. 1–14.

Hartmann, B. 2015. Die hölzernen Things: Commodities in Cultural Schreibtafeln im Imperium Romanum – ein Perspective. Cambridge: Cambridge Inventar. In: M. Scholz & M. Horster, eds. University Press, pp. 64–91. Lesen und Schreiben in den römischen Kühlborn, J.-S. 1992. Das Römerlager in Provinzen.Mainz:Römisch-Germanisches Oberaden III. Münster: Aschendorff. Zentralmuseum, pp. 43–58. Küster, H. 1994. The Economic Use of Abies Hather, J.G. 2000. The Identification of the Wood as Timber in Central Europe Northern European Woods: A Guide for during Roman times. Vegetation History Archaeologists and Conservators. London: and Archaeobotany,3:25–32. Archetype Press. Linnane, J.L. & Kinsella, J. 2009. Military Hicks, D. & Beaudry, M.C. 2010. Lords and Defensive Beginnings: A Introduction: Material Culture Studies, a Preliminary Assessment of the Social Role Reactionary View. In: D. Hicks & M.C. of an Impressive Rath at Baronstown. In: Beaudry, eds. The Oxford Handbook of M.B. Deevy & D. Murphy, eds. Places Material Culture Studies. Oxford: Oxford Along the Way: First Findings on the M3. University Press, pp. 1–21. Dublin: National Roads Authority, pp. Hodder, I. 2014. The Entanglements of 101–22. Humans and Things: A Long-Term Lodwick, L.A. 2017. Evergreen Plants in Roman View. New Literary History, 45: 19–36. Britain and Beyond: Movement, Meaning, Hodder, I. & Lucas, G. 2017. The and Materiality. Britannia, 48: 135–73. Symmetries and Asymmetries of Human– https://doi.org/10.1017/S0068113X17000101 Thing Relations: A Dialogue. Mair, C. 2013. The History of the Incorporation Archaeological Dialogues, 24: 119–37. of Coopers of Glasgow. Glasgow: Neil Hopf, M. 1967. Einige Bemerkungen zu Wilson. römerzeitlichen Fässern. Jahrbuch des Malafouris, L. 2013. How Things Shape the römisch-germanischen Zentralmuseums Mind: A Theory of Material Engagement. Mainz, 14, 212–216. Cambridge (MA): MIT Press. Horn, J. 2015. Tankards of the British Iron Marchand, T.H.J. 2010. Embodied Cognition Age. Proceedings of the Prehistoric Society, and Communication: Studies with British 81: 311–41. Fine Woodworkers. Journal of the Royal IAWA Commitee 1989. IAWA List of Anthropological Institute, 16, S100–S120. Microscopic Features for Softwood https://doi.org/10.1111/j.1467-9655.2010. Identification. International Association of 01612.x Wood Anatomists (IAWA) Journal, 25: Marlière, E. 2001. Le tonneau en Gaule 1–25. romaine. Gallia, 58: 181–201. Joy, J. 2009. Reinvigorating Object Biography: Marlière, E. 2002. L’outre et le tonneau dans Reproducing the Drama of Object Lives. l’Occident romain (Monographies World Archaeology, 41: 540–56. Instrumentum). Montagnac: Monique Joyce, R.A. & Gillespie, S.D. eds. 2015. Mergoil. Things in Motion: Object Itineraries in Marlière, E. 2003. Tonneaux et amphores à Anthropological Practice. Santa Fe (AZ): Vindolanda: contribution à la connaissance School for Advanced Research. de l’approvisionnement des troupes Kilby, K. 1971. The Cooper and his Craft. stationnées sur la frontière Nord de London: J. Baker. l’Empire. In: A. Birley, ed. Vindolanda Knappett, C. 2005. The Affordance of Research Report: The Excavation of 2001– Things: A Post-Gibsonian Perspective on 2002, Volume 1. Hexham: Vindolanda the Relationality of Mind and Matter. In: Trust, pp. 125–79. E. Demarrais, C. Gosden & C. Renfrew, Marlière, E. 2014. Les campagnes militaires et eds. Rethinking Materiality: The l’expansion de l’usage du tonneau dans Engagement of Mind with the Material l’Empire romain. In: M. Perrot & J. World. Cambridge: MacDonald Institute Pérard, eds. De la cave au vin: une fruc- for Archaeological Research, pp. 43–51. tueuse alliance. Dijon: Université de Kopytoff, I. 1986. The Cultural Biography of Bourgogne, pp. 47–61. Things: Commoditization as Process. In: Marlière, E. & Costa, J.T. 2005. Tonneaux et A. Appadurai, ed. The Social Life of amphores à Vindolanda: contribution à la

connaissance de l’approvisionnement des Log Quality and Prediction of Sawing troupes stationnées sur le mur d’Hadrien Yield in Oak Wood (Quercus robur). (II). In: A. Birley & J. Blake, eds. Annals of Forest Science, 70: 695–706. Vindolanda: The Excavation of 2003/2004. Ross, L.A. 1985. 16th-Century Spanish Hexham: Vindolanda Trust, pp. 214–36. Basque Coopering. Historical Archaeology, Mauri, A., De Rigo, D. & Caudullo, G. 2016. 19: 1–31. Abies alba in Europe: Distribution, Rubio-Campillo, X., Coto-Sarmiento, M., Habitat, Usage, and Threats. In: J. San- Pérez-Gonzalez, J. & Rodríguez, J.R. Miguel-Ayanz, D. De Rigo, G. Caudullo, 2017. Bayesian Analysis and Free Market T. Houston Durrant. & A. Mauri, eds. Trade within the Roman Empire. European Atlas of Forest Tree Species. Antiquity, 91: 1241–52. Luxembourg: European Commission, pp. Rule, M. & Monaghan, J. 1993. A Gallo- 48–49. Roman Trading Vessel from Guernsey: The Morris, C.A. 2000. Wood and Woodworking in Excavation and Recovery of a Third Anglo-Scandinavian and Medieval York Century Shipwreck (Guernsey Museum (The Archaeology of York 17(13): The Monograph No. 5). Guernsey: Guernsey Small Finds). York: Council for British Museum & Galleries. Archaeology. Saedlou, N. & Dupéron, M. 2007. Etude Nakagawa, T., De Beaulieu, J.-L. & Kitagawa, xylologique et typologique des tablettes à H. 2000. Pollen-derived History of écriture antiques en bois à partir des Timber Exploitation from the Roman découvertes faites à Saintes (Charente- Period Onwards in the Romanche Valley, Maritime). In: J.-L. Dupouey, E. Central French Alps. Vegetation History Danbrine, C. Dardignac & M. George- and Archaeobotany,9:85–89. Leriy, eds. La mémoire des forêts: actes du Olsen, B. 2010. In Defense of Things: colloque Forêt, archéologie et environnement, Archaeology and the Ontology of Objects, 14–16 décembre 2004. Paris: Office Lanham (MD): AltaMira. national des forêts, Institut national de la Padley, T.G. 1991. The Wooden Objects. In: recherche agronomique & Nancy: T. Padley & C. Winterbottom. The Direction régionale des affaires culturelles Wooden, Leather and Bone Objects from de Lorraine. Castle Street, Carlisle Excavations Sands, R. & Hather, J.G. in prep. Wooden (Cumberland and Westmoreland Small Finds from Vindolanda. Vindolanda Antiquarian and Archaeological Society Trust. Research Series 5). Kendal: Cumberland Sands, R. & Horn, J.A. 2017. Bring me Three and Westmorland Antiquarian and Large Beers: Wooden Tankards at Roman Archaeological Society, pp. 203–27. Vindolanda. Oxford Journal of Archaeology, Pearce, J. 2004. Archaeology, Writing Tablets 36: 503–15. and Literacy in Roman Britain. Gallia, 61: Schiffer, M.B. 1972. Archaeological Context 43–51. and Systemic Context. American Antiquity, Pollard, A.M., Bray, P., Gosden, C., Wilson, 37: 156–65. A. & Hamerow, H. 2015. Characterising Schiffer, M.B. 2010. Behavioural Archaeology: Copper-based Metals in Britain in the Principles and Practice. London and First Millennium AD: A Preliminary New York: Routledge. Quantification of Metal Flow and Schweingruber, F.H. 1990. Microscopic Wood Recycling. Antiquity, 89: 697–713. Anatomy: Structural Variability of Stems Reis, A. 1997. Roman Barrel Finds from and Twigs in Recent and Subfossil Woods Vindolanda (unpublished MA dissertation, from Central Europe. Birmensdorf: Durham University). Eidgenössische Forschungsanstalt. Ribereau-Gayon, P. 1994. The Barrel and the Sciallano, M. 1993. L’art du tonnelier. Istres: Wine: Scientific Mastering of a Traditional Musée d’Istres. Know-how. Seguin Moreau Cooperage. Napa Sealey, P.R. 2009. New Light on the Wine (CA): Enology Institute of the University of Trade with Julio-Claudian Britain. Bordeaux, Seguin Moreau USA. Britannia, 40: 1–40. Riesco Muñoz, G., Remacha Gete, A. & Stewart, K. 2016. Identifying the Wooden Gasalla Regueiro, M. 2013. Variation in Tablets to Species. In: R.S.O. Tomlin, ed.

Roman London’s First Voices: Writing medieval Ireland, Britain, and Europe. He Tablets from the Bloomberg Excavations, researches aspects of the application of – 2010 2014 (MOLA Monograph 72). technology to archaeology and artefact London: Museum of London, p. 6. Swift, E. 2012. Object Biography, Re-use and analysis and is a lecturer at UCD School Recycling in the Late to Post-Roman of Archaeology, Dublin. Transition Period and Beyond: Rings Made from Romano-British Bracelets. Address: Rob Sands, University College – Britannia, 43: 167 215. Dublin, School of Archaeology, Newman Tegtmeier, U. 2016. Holzobjekte und Holzhandwerk im römischen Köln: Building, Belfield, Dublin 4, Ireland. Archäologie Nord-Süd Stadtbahn Köln. [email: [email protected]]. ORCID: Köln: Römisch Germanisches Museum. 0000-0003-3458-0229. Tomlin, R.S.O. 2016. Roman London’sFirst Voices: Writing Tablets from the Bloomberg Excavations, 2010–14 (MOLA Monograph 72). London: Museum of London. Elise Marlière was born in Abbeville Ulbert, G. 1959. Römische Holzfässer aus (Somme, France). In 2000, she obtained Regensburg. Bayerische Vorgeschichtsblätter, her doctorate in Sciences of Antiquity 24: 6–29. (University of Lille III), which explored Verboven, K. 2011. Professional Collegia: the use of amphorae, leather bags, and Guilds or Social Clubs? Ancient Society, barrels in the Roman world, particularly in 41: 187–95. Viérin, J. & Léva, C. 1961. Un puits à the northwest of Gaul, and led her to tonneau romain avec sigles et graffiti à study the supply of the Roman army. She Harelbeke. Latomus, 20: 759–84. has taught at Poitiers University and Wilson, A. 2009. Approaches to Quantifying worked with the ‘Gaule’ team at the Roman Trade. In: A.W.A Bowman & A. CNRS (UMR 7041, Paris X). She now Wilson, eds. Quantifying the Roman Economy. Oxford: Oxford University Press, lives in Ibiza (Spain), where she co- pp. 213–49. directs, with J. Torres Costa, Antiquarium, a company dedicated to rescue archaeology and protection of the ’ BIOGRAPHICAL NOTES island s heritage. She has published in French, Spanish, and Catalan and is pre- Rob Sands has an undergraduate degree paring publications on the Budapest and from the Institute of Archaeology, Menföcsanak barrels (Hungary) and the University of London, and a PhD from amphorae and barrels from Vindolanda. the University of Edinburgh, analysing toolmarks on woodwork from Iron Age Address: Elise Marlière, Avenida Isidoro lake-dwellings in Scotland. His research Macabich 19, 3 7ª, 07820 Sant Antoni de interests centre on wetland archaeology Portmany,Ibiza,BalearicIslands,Spain. and the examination of organic artefacts in [email: [email protected]]. later prehistoric, Roman, and early ORCID: 0000-0002-5036-3278.

Production, réparation, réutilisation, adaptation et recyclage : les multiples biographies d’un tonneau romain

Au cours des premières décennies du second siècle apr. J.-C., alors que l’empire romain était à son apogée, les tonneaux en bois étaient un élément essentiel d’un réseau de commerce à la base d’un système économique complexe et étendu. Cependant ces objets ne sont normalement pas conservés en archéologie

et peu de sites ont livré des ensembles suffisamment grands et provenant de plusieurs phases. Quoique relativement rares, des trouvailles individuelles et des ensembles ont été recueillis de façon suffisamment régulière pour permettre un examen de la production et de l’usage des tonneaux à l’époque romaine. Ces objets peuvent avoir une biographie culturelle complexe, allant de leur lieu de production jusqu’à leur déposition finale. Les recherches menées récemment et antérieurement à Vindolanda, un fort romain du nord de la Grande-Bretagne aux confins de l’empire romain, fournissent un cadre de réflexion sur ces objets et leurs biographies. Ici nous mettons l’accent sur les indices qui démontrent une pratique réitérée, peut-être habituellement répétée, d’adaptation et de recyclage perceptible dans ce matériel. Translation by Madeleine Hummler

Mots-clés: époque romaine, tonnellerie, Abies alba, Picea abies, Larix decidua, biographie d’objets recyclage

Herstellung, Reparatur, Wiederverwendung, Anpassung und Wiederverwertung: die vielfältigen Biografien von römischen Fässern

In den ersten Jahrzehnten des 2. Jahrhunderts n. Chr., als das römische Reich am ausgedehntesten war, spielten die Holzfässer eine wichtige Rolle in einem Handelsnetzwerk, das eine Komplexe und weitrei- chende förderte. Diese Gegenstände sind aber normalerweise in den archäologischen Befunden nicht erhalten und es gibt nur wenige Fundstellen mit umfangreichen mehrphasigen Sammlungen. Obschon sie recht selten vorkommen, sind Einzelfunde und Sammlungen von Fässern regelmäßig genug gefunden worden, um eine Untersuchung der Herstellung und Gebrauchs von Fässern in der Römerzeit zu erlau- ben. Diese Gegenstände können eine komplexe kulturelle Biografie haben, die vom ursprünglichen Herstellungsort bis zur endgültigen Deponierung reichen. Laufende und abgeschlossene Untersuchungen in Vindolanda, ein Kastell in Nordbritannien am Rande des römischen Reiches, geben uns die Gelegenheit, über diese Gegenstände und deren Biografien nachzudenken. Hier erwägen wir besonders, ob man in diesem Material eine mehrfache, vielleicht gewöhnlich wiederholte Praxis der Anpassung und Wiederverwertung nachweisen kann. Translation by Madeleine Hummler

Stichworte: Römerzeit, Küferei, Abies alba, Picea abies, Larix decidua, Biografie eines Gegenstandes, Wiederverwertung