Étude de trois grands tonneaux mis au jour à Reims/Durocortorum (Marne) : le savoir-faire des tonneliers antiques

• Published in: Gallia Vol. 77; no. 2; pp. 123 - 155
• Main Authors: Mille, Pierre, Rollet, Philippe
• Format: Journal Article
• Language: English
• Published: CNRS Éditions 31.12.2020
• ISSN: 0016-4119; 2109-9588
• DOI: 10.4000/gallia.5242

Preventive archaeological excavations carried out in Reims have provided the opportunity to uncover and study for the first time the antique deposits on the right bank of the Vesle river along a length of about fifty metres. The results obtained from these investigations made it possible to trace the evolution of this urban sector according to four phases ranging from the beginning of the Current Era to the 5th c. AD. During phase 1 and as early as the second half of the reign of Augustus initial traces of settlement appear, but major construction work along the river-side was not completed before 15/20 AD. During phase 2, which starts around 85-90 AD, a grid of streets bordered by wooden gutters was laid out. These gutters drained sewage and rainwater towards the first quay on the right bank. At the beginning of phase 3, from the first half of the 1st c. AD on, this sector was completely incorporated into the city. The area hosted craft activities and a new quay was built. From 306-310 AD (phase 4) on a monumental building was erected. It contained a series of adjoining furnaces, but the function of this imposing complex, which was destroyed at the end of the 4th c. AD, has not yet been identified. Apparently the area was deserted for a long period of time, because indicators of activities carried out in the sector can be dated only to the 14th c. On a strip near the right bank of the Vesle river three water wells (Pts 378, Pts 354 and Pts 417) were excavated. They were related to the craft activities of phase 3 (between the beginning of the 2nd c. and the beginning of the 3rd c. AD). The casings of these wells were particular in that vats were reused. These three large barrels were in an outstanding state of preservation which made it possible to carry out a very faithful reconstruction. Each of these barrels was made up of 22 to 25 tapered, 2-m-long staves and contained between 1,000 and 1,200 litres of wine, as was evidenced by analyses of the deposits of malic acid and tartaric acid on their walls, which are both indicators of alcoholic fermentation. Their manufacture, their use as wine containers and their reuse have been dated using dendrochronology and the study of pottery contained in the fillings, to between the second half of the 1st c. AD and the very beginning of the 4th c. AD. Technical analysis of the three vats made it possible to draw up a precise list of the crafts involved in their fabrication, detailing for each of these their respective skills and tools. Some of these crafts were linked to sylviculture, others were carried out in urban workshops, and the fact that they were separated by great distances did not hinder their collaboration. Use wear analysis was based on the traces left by the tools, still recognisable on the surfaces of the staves. It was completed by the anatomical identification of the wooden pieces and their original position in the log, which makes it possible to identify in more detail the operational sequence. The staves of the three casks were cut preferentially on a radial section, i.e. a priori by splitting off from large logs of fir (Abies alba) logs. Forestry activity, which consists of trimming the planks to be transformed into staves according to the lengths, widths and thicknesses ordered by the cooper, has been carried out by cleaving craftsmen since Antiquity up to the present day. The planks are then dried for a length of time that depends on the thickness of the planks and the nature of the wood. The cooper, or cūparius chooses a batch of planks to build his barrel. Each stave is then subjected to a series of very precise shaping operations. For the backing the cooper uses a doloire, an axe with a large blade. This tool left large facets on the rounded back of the staves, often at an angle of 45° to the vertical plane and organised along three longitudinal planes. To provide the staves with a pointed end the cooper planes the edges of both ends of the planks. These rigorously flat and properly dipping edges were skilfully straightened using a plane with two handles. At the staves’ end the thickness of the top part was evened out using a specific adze with a curved blade. The croze grooves, thin transverse grooves into which the heads were fitted, were cut using a croze (another specific cooper’s tool). The hoops are not preserved equally well on all three vats; there were between two and four series of adjoining hoops of varying width placed on either side of the bilge. All the hoops were made from hazelnut saplings (Corylus avellana) with a semi-circular section, which were split and cut using a billhook (barrel hoop maker’s tool), then bent and tied. The analysis of one of the cords that served as a tie, stemming from barrel 354, has shown that it was probably made from esparto grass (Lygeum spartum), a Poaceae species found throughout the Mediterranean area. In order to meet the orders of the cooper’s workshops the hoop makers –another forest-related activity– had to obtain their supplies from the cord-makers, among other peoples. Remains of pitch, an efficient sealing material, were visible on all the internal surfaces of the staves, up to the level of the crozes. The pitch coating, which ensured the sealing of the three vats, was up to 0,50 cm thick at some points. During Antiquity pitch was produced by pyrogenation (distillation) of pine wood (Pinus sp.) in large jars. This production was carried out in forests in dedicated areas, by specialised craftsmen whose names have not been recorded (during Antiquity this is a recurrent observation for all forestry crafts which can be explained by the low social status of this small corporation of craftsmen). An ad hoc amount of pitch was liquefied in a fire and then brushed on the interior of the barrels while the heated cask was rolled vigorously. The barrels were sealed by the coopers in the workshops. On the wooden surface of the three barrels over 45 branded or stamped marks were preserved using various signacula and adzes with a marking punch, often tria nomina, sometimes numbers. The essential contribution of the analysis of these marks lies in the fact that it is possible to make a clear distinction between struck marks and branded marks, as the positioning of each of these apparently depends on formal and recurrent practices that vary for each craftsman who placed them on the barrel. The marks found in Reims, in addition to those on other European vats –nearly 200 examples are currently listed (most of which were found along the Roman Limes)– confirm a long and complex operational sequence, which involved very numerous specialised craftsmen who signed their work. First of all, the cleaving craftsmen left marks struck using adzes with a marking punch. These marks are engraved in the fresh wood and are readable and durable, for example the marks CVPRC or CVPRO identified on vat 354, or the doubles: A˙AP∙VD on vat 417. Stave makers used these marks, specific to their craft, to mark their work with their name or that of the workshop owner, thus fixing its quality and cost, doubtlessly during organised sales. The cooper who assembles the staves in his workshop leaves another type of mark, called a reference mark, which is used on the workbench to ensure the best possible juxtaposition of the staves. The series of numbers I to XXII printed on the staves of vat 354 illustrate this. The assembling was entrusted to a master craftsman, as the strength, sealing and durability of the barrel depended on him. Once the barrel was assembled, the cooper branded his mark on it with a hot iron. These large branded cartouches appear as a highlight on the bottom of the barrels, often positioned transversely and encompassing two staves, as well as directly above the bung hole. In Switzerland several types are known, for example the one discovered in Eschenz: C.ANTSPEN͡D. In antique times the merchant was the owner of the barrels and he was also the one who supplied them to the winegrower to fill them with the wine of the year. The branded marks of the wine merchants regularly appear as seals on the bungs at the bilge of the barrels. Once the barrel was filled, it was mandatory to add a mark. This is the case of the COSAT branded mark of vat 378 of the Reims site, which was repeated three times, with the seals half on the bung and half on the vents. Certain branded stamps placed on the bilge near the bungs very explicitly designate the receiver and can be easily distinguished from the stamps of the wine merchants. Once the barrels were filled and sealed, the wine merchant handed them over to a conveyor and nautae to transport these heavy loads. A rich stone inventory from all over Gaul and Germania provides precise pictures of these land vehicles and barge transports along the major river routes of the Rhone and the Rhine at that time. On the bilges of some large barrels there are also abundant branded, more rarely stamped, short inscriptions –sometimes a single letter, or even two or three, often applied on or next to the bung stave. This is the case for the small, short inscriptions CTM and CON found on vat 354. They are clearly distinguishable from the other cartouches mentioned above, but it is not clear whether they are those of conveyors or controllers. It is assumed that numerous sworn agents who were responsible for controlling sales and collecting taxes prior to the barrel reform affixed this type of short mark. More anecdotally, there are also numerous graffiti on the bottom of the barrels made with a dry point. Many of these refer to private individuals, who in this way claimed ownership of the cask, but several others followed by the word fecit clearly name the winegrower who produced the content. More rarely, these graffiti indicate the type of wine and the quantity shipped. While the vats reserved for the armies (a different supply chain) were filled with acetum (sour wine) used for the preparation of posca (t