Navigation – Plan du site

AccueilNuméros43-1The biggest Opus doliare producti...

The biggest Opus doliare production in Narbonese Gaul revealed by archaeometry (first to second centuries A.D.)

La plus grande production d’Opus doliare de Gaule Narbonnaise révélé par l’archéométrie (ier-iie s. apr. J.-C.)
Charlotte Carrato, Verónica Martínez Ferreras, Jean-Marie Dautria et Michèle Bois
p. 69-82

Résumés

Cet article présente une synthèse des récentes recherches menées sur les dolia (grandes jarres de stockage) du territoire de l’ancienne province de Gaule narbonnaise. Il vise à mettre en lumière un phénomène méconnu, à savoir la production et la diffusion des dolia à l’époque romaine par le biais des données archéométriques. Cette étude démontre comment le croisement des données archéométriques, géologiques et archéologiques ont permis de révéler l’existence de l’une des plus grandes zones de production de dolia de Gaule Narbonnaise.

Haut de page

Notes de la rédaction

rec. apr. 2018; acc; may 2019.

Notes de l’auteur

This work is an extension of the PhD thesis published in 2017 in the collection Mémoire d’Ausonius-Bordeaux (Carrato 2017).

Texte intégral

The study was funded by LaBex ARCHIMEDE-Montpellier as part of the Investissement d’Avenir programme ANR-11-LABX-0032-01. The research is part of Axis 3, Economy and environment LabEx and benefitted from research funding for the geochemical analyses and the English translation. The corresponding author specially thanks to Dr Stéphane Mauné (CNRS Research Director, UMR5140-ASM-Montpellier), who initiated her research on dolia, for his revisions and corrections. She is also grateful to André Tchernia for reading this text and suggesting improvements. She also wish to thank Oriane Bourgeon and Séverine Corbeel (PhD students UMR5140-ASM-Montpellier/LabEx ARCHIMEDE-Montpellier) for their bibliographic guidance.

(Michèle Bois.) My thanks to Stéphane Mauné (CNRS Research Director, UMR5140-ASM-Montpellier), who initiated my research on dolia, for his revisions and corrections. I am also grateful to André Tchernia for reading this text and suggesting improvements. I also wish to thank Oriane Bourgeon and Séverine Corbeel (PhD students UMR5140-ASM-Montpellier/LabEx ARCHIMEDE-Montpellier) for their bibliographic guidance.

1. Introduction

1The dolium is a ceramic storage jar of Greek and Etruscan origin, and is the largest container ever used in the ancient world. It appeared in Gaul during the fourth century B.C., became increasingly widespread until the first century B.C., and then flourished during the Early Empire. At that time, jars were installed in large storehouses and their capacity ranged from 600 to 2,500 l, ensuring the storage of significant quantities of wine and olive oil (Fig. 1).

Figure 1: Dolium from the Bernardins villa in Montélier (Drôme) (capacity: 2,463 l) conserved in the Museum of art and archaeology of Valence / Dolium provenant de la villa des Bernardins à Montélier (Drôme) (capacité : 2 463 l) conservé au musée d’art et d’archéologie de Valence

Figure 1: Dolium from the Bernardins villa in Montélier (Drôme) (capacity: 2,463 l) conserved in the Museum of art and archaeology of Valence / Dolium provenant de la villa des Bernardins à Montélier (Drôme) (capacité : 2 463 l) conservé au musée d’art et d’archéologie de Valence

Cl. C. Carrato.

  • 1 Known as In dolia fugularem artem discere by Plato, Gorgias, 514e and Lachès, 187b, or Erasmus, Les (...)
  • 2 See in particular Descemet, 1880 ; Steinby, 1993 ; Gasperoni, 2003 ; 2005 ; Lazzeretti and Pallecch (...)

2Judging by early sources, the difficulty in manufacturing these large terracotta jars was proverbial1. Contrary to long-standing common belief, each stage, from the preparation of the clay, to the difficult stage of manually mounting coils, to firing, required the savoir-faire of the most experienced potters, as suggested by ethnographic parallels (Checa, 1901 ; Romero Vidal and Cabasa Calpe, 1999 ; Giannopoulou, 2010 ; Vaz Pinto, 1997). A significant number of dolia were discovered at early agricultural sites from the Roman period in Narbonese Gaul, following the expansion of viticulture and olive growing during the course of the first century A.D. (Brun, 2005 ; Brun and Laubenheimer, 2001), indicating that the elders adapted the production of these containers to economic demands. Paradoxically, production sites and trade modes are still poorly known. For the whole Mediterranean region, the only clearly evidenced production workshop is the Saint-Bézard villa in Aspiran (Hérault, France) (Mauné, et al., 2006) and the Ermedàs workshop in Cornellà del Terri (Catalogna, Spain) (Tremoleda, et al., 2018). Others production sites are only indirectly known through epigraphic stamps2, which provide information concerning the status of those involved in this specialized production, but are of limited use in research on workshop localisation and production diffusion modes.

3In order to address the paucity of archaeological data, the dolia from wine and oil exploitations in Narbonese Gaul were studied by one of the authors as part of a PhD thesis submitted in 2014 (Carrato, 2017, 158-164). The corpus of this study has a total of six hundred and fifty dolia from seventy three sites located in the coastal part of Narbonese Gaul, extending from the Pyrénées-Orientales to the Alpes-Maritimes, but also along the Rhône Valley to Lyon (Fig. 2). The macroscopic examination allowed classifying the specimens in different fabrics, which were more accurately defined through the archaeometric characterisation of a number of selected samples. The methodology adopted aimed to combine all of the available approaches, namely macroscopic classification, and the examination of the chemical, mineralogical and petrographic compositions in order to establish their provenance and the main technological processes of manufacture. Thus, the petrographic examination of one hundred and twenty-five samples selected from the initial corpus then led to distinguish forty-five regional paste types. The matching between regional groups required chemical analysis on thirty-six samples which appeared to be representative of four main petrographic fabrics. The results of this study allowed distinguishing the most common dolia pastes found in Narbonese Gaul dated between the first century and the middle of the second century A.D., to make inferences on their provenance and the location of the possible production centres, to evaluate their frequency and the main areas of distribution. The chronology and the extension of the area of influence of these workshops points to two main production and distribution systems. One includes thirty-eight workshops with micro-regional diffusion and the other refers to three workshops, or groups of workshops with regional, or even provincial diffusion.

Figure 2: Distribution map of the sites in the studied corpus (H. Bohbot and C. Carrato) / Carte de répartition des sites du corpus d’étude (H. Bohbot et C. Carrato)

Figure 2: Distribution map of the sites in the studied corpus (H. Bohbot and C. Carrato) / Carte de répartition des sites du corpus d’étude (H. Bohbot et C. Carrato)

4Among the latter, a group of pastes (called group 1), with predominant calcite inclusions is widespread in Narbonese Gaul. During the period under study (first-second centuries A.D.), these pastes were used for one dolium out of three in certain regions. This group also presents the most extensive diffusion zone, extending over nearly 400 km, from Lyon to Narbonne, through Arles and the outskirts of Marseille, corresponding to almost the whole Narbonese province. Due to the abundance of this fabric in Roman sites, its homogeneity and the extent of its diffusion zone, it seems to reflect an unusual and unprecedented production system in the province. The present study focuses on the complete archaeometric characterisation of the dolia from different consumption centres attributed to group 1. The aim of the present study is to thoroughly evaluate the degree of homogeneity in the composition of these vessels, to determine the technological processes involved in their manufacture and to identify their provenance (the location of the production area).

2. Sampling and methods

Macroscopic characterization

5The macroscopic examination was carried out with the naked eye or with a magnifying glass on fresh breaks

6Considering the size of the inclusions in the dolia pastes, the macroscopic examination is a useful technique for an initial classification of the specimens under study. They consist of coarse to very coarse fabrics, although they differ in the composition and characteristics of the clayey matrix and the non-plastic inclusions. Indeed, the corpus of dolia studied by the main author (Carrato, 2017) includes specimens manufactured using both, Ca-rich clays or Fe-rich clays. The observation of the matrix colour mainly reveals oxidizing firing. Nevertheless, the central area of the thick walls of some individuals is brown because it was less affected by the firing temperature than the internal and external surfaces of the jars. In other cases, the core of the ceramic walls is greyish, suggesting changes in the atmosphere of the kiln during the firing process. In addition, the examination of complete specimens indicates that a single dolium can present different surface colours. This different colouring is generally linked to the position of the container in the kiln and cannot consequently be taken into consideration for macroscopic identification. The examination of the non-plastic inclusions was a more convincing discriminating marker. In general, they consist of coarse to very coarse grains of different nature that were added as temper to the clayey raw material. Most of the individuals of the corpus appear to have been tempered with crushed calcite, while coarse to very coarse sand grains from a plutonic and or volcanic geological complex were added to the rest of the fabrics..

7The macroscopic group 1 comprises around two hundred dolia from the main corpus. All of them exhibit a dark-orange to dark-red matrix, rarely grey-black. The non-plastic material is very abundant and poorly sorted. It is mainly composed of fine, coarse and very coarse white inclusions, usually quadrangular and angular, consisting of monocrystalline or polycrystalline crushed calcite. They appear associated with rarer rounded wine-coloured or black grains which probably correspond to ferruginous nodules (Fig. 3).

Figure 3: Macroscopic views of fragments of group 1 paste (1,5x) / Vues macroscopiques de fragments de pâte du groupe (1,5x)

Figure 3: Macroscopic views of fragments of group 1 paste (1,5x) / Vues macroscopiques de fragments de pâte du groupe (1,5x)

The numbers refer to the sites listed in Fig. 2. / Les numéros renvoient aux sites répertoriés dans la Fig. 2.

Cl. C. Carrato.

Archaeometric characterization

  • 3 DOL1, DOL2 : La Nautique, Narbonne (Aude) ; DOL3, DOL5 : La Barque, Allan (Drôme) ; DOL4, DOL7 : Le (...)

8Complementary analyses were made on twelve samples attributed to group 1 in order to confirm and refine the macroscopic determination. Those specimens were coded DOL1 to DOL12 (Carrato, 2017, 223, fig.114)3. They come from nine wine exploitations located in the Hérault, Aude and Drôme areas, and they are dated between the first century and the middle of the second century A.D. The investigation included the petrographic composition of the selected samples and the chemical and mineralogical composition of the same (powdered) specimens. The analyses were carried out at the CCiT-UB laboratory of the University of Barcelona, following the procedures already described (Martinez Ferreras, et al. 2013), and at the Geoscience laboratory of the University of Montpellier II.

9The petrographic composition was examined through thin-section optical microscopy (OM), using an Olympus BX43F polarising microscope; measurements, evaluations, and photographs were taken using an Olympus DP73-WDR digital camera and the Stream Basic image analysis software.

10The chemical composition was investigated through wavelength dispersive X-ray fluorescence (WD-XRF), using a Panalytical-Axios PW 4400/40 spectrometer. This allowed for the identification of 29 major, minor and trace elements, including Fe2O3 (as total Fe), Al2O3, MnO, P2O5, TiO2, MgO, CaO, Na2O, K2O, SiO2, Ba, Rb, Th, Nb, Pb, Zr, Y, Sr, Ce, Ga, V, Zn, Cu, Ni, Cr, Mo, Sn, Co and W. Of these, Mo and Sn were discarded due to their low counting statistics, while Co and W due to a possible contamination from the tungsten carbide cell of the mill used in the preparation of the samples. Glass beads were prepared for the determination of major and minor elements, using 0.3 g of specimen. The concentration of the trace elements was measured through powder pills prepared using 5 g of specimen. The quantification of the elements was obtained by using a calibration line performed with 60 International Geological Standards. The loss on ignition (LOI) was calculated from calcinations of 0.3 g of dried specimen, heated at 950˚C for 3 hours.

11The mineralogical composition was determined by powder X-ray diffraction (XRD), using a PANalytical X’Pert PRO MPD powder diffractometer. Crystalline phases were investigated using the software HighScore Plus by PANalytical.

3. Results

Petrographic characterization

  • 4 The inclusions in the fine fraction correspond to the smallest grains ( 0.125 mm) and are abundant (...)

12Thin section analysis was carried out on the twelve selected samples in order to confirm and refine the macroscopic determination. The petrographic fabrics were defined according to the characteristics of the matrix (groundmass) and the microstructure (micromass). We paid particular attention to the identification and description of the aplastic inclusions (frequency, size, shape, distribution, and type). Each mineralogical description follows a bipartite presentation itemizing the minerals in the fine fraction, then the minerals in the coarse fraction4.

13The petrographic composition confirms that all the specimens analysed consist of calcite-tempered dolia, although some differences in the characteristics of the groundmass and the presence or absence of specific non-plastic inclusions have been recognised (Fig. 4a). Most of the specimens exhibit a Fe-rich, heterogeneous and low vitrified matrix. The fine fraction is abundant and is mainly composed of small crystals of quartz together with common phyllosilicates and opaque minerals. Some aggregates of a different clayey material, also containing small crystals of quartz, appear combined with the main clay, suggesting that different clayey raw materials were probably mixed. The coarse fraction is mainly composed of angular and rhombohedral shaped mono-crystalline calcite (≤ 10 mm) followed by subrounded to rounded ferruginous grains (≤ 8 mm), which could be considered fine-grained sedimentary rocks with Fe-rich matrix. Subangular to subrounded monocrystalline quartz (< 1 mm) is frequent; subrounded fragments of metamorphic rocks (principally quartzite and quartz-mica schist, < 1 mm) are common, and lames of muscovite (< 0.5 mm) are rare. DOL8 presents a very similar clay matrix and the non-plastic inclusions correspond to the same lithic elements already described (Fig. 4b). Nevertheless, the crystals of quartz and metamorphic rocks are more abundant. In addition, nodules of monocrystalline (micrite) calcite are common. The dolium DOL3 exhibits the same non-plastic inclusions although it differs in the characteristics of the matrix (Fig. 4c). The latter is rather heterogeneous and contains large and elongated streaks of very pure clay mixed with a more abundant clayey material, darker in colour. The optical activity of the matrix is very high suggesting a low firing temperature. The fine fraction is less frequent and finer than in the previous fabrics. It comprises very small lames of phyllosilicates, very fine grains of quartz and opaque minerals. Angular and rhombohedral mono-crystalline calcite (≤ 4 mm) is prevalent in the coarse fraction, together with some nodules of microcrystalline (micrite) calcite. Ferruginous nodules and subangular to subrounded monocrystalline quartz (< 1 mm) are frequent, while quartzite and quartz-mica schist are common.

Figure 4: Microphotographs of thin sections taken under plane polars (PPL) and crossed polars (XPL) at the same magnification (20x); a) representative individuals of the main fabric identified; b) the dolium DOL8; c) the dolium DOL3 / Microphotographies des lames minces en lumière polarisante non analysée (LPNA) et en lumière polarisante analysée (LPA) au même grossissement (20x) ; a. Individus représentatifs de la même fabrique ; b. le dolium DOL 8 ; c) le dolium DOL3 (V. Martínez Ferreras)

Figure 4: Microphotographs of thin sections taken under plane polars (PPL) and crossed polars (XPL) at the same magnification (20x); a) representative individuals of the main fabric identified; b) the dolium DOL8; c) the dolium DOL3 / Microphotographies des lames minces en lumière polarisante non analysée (LPNA) et en lumière polarisante analysée (LPA) au même grossissement (20x) ; a. Individus représentatifs de la même fabrique ; b. le dolium DOL 8 ; c) le dolium DOL3 (V. Martínez Ferreras)

14Apart from the differences detected, all the dolia analysed were manufactured using the same type of non-plastic inclusions. Their composition and characteristics suggest that raw materials of different geological origin were mixed to obtain specific coarse clayey pastes. Indeed, the shape and composition of the fine fraction denotes the use of a Fe-rich alluvial sediment taken from a river or river banks. The subangular to subrounded crystals of quartz and metamorphic rocks and the ferruginous nodules, usually ≤ 8 mm grain-sized, could be naturally included in the alluvial sediments. However, the shape and the variety of size that characterize the crystals of calcite suggest that they were added by crushing a calcareous rock from a karstic source which forms mainly in limestone regions.

Geochemical characterization

15The geographic range of the group 1 specimens raised questions as to the homogeneity of these vessels. In addition to the data provided by the petrographic analysis, the examination of the geochemical characterization is required to evaluate the compositional variation among the dolia classified in group 1. In the scope of studies on the provenance of archaeological ceramics, the geochemical characterization is especially useful when the reference chemical groups of the production centres are known. Since the dolia examined herein were recovered in consumption centres and no analytical database of the same materials is available, the aim of this analysis is to evaluate the dis(similarity) of the chemical composition of the jars attributed to group 1 in order to test the coherency of the macroscopic and petrographic fabrics previously defined.

16The normalised chemical composition of the twelve samples analysed by wavelength dispersive X-ray fluorescence (WD-XRF) is given in Figure 5a. It reveals that all the vessels consist of calcareous ceramics (CaO from 10 to 19 wt%), except DOL5 that is border calcareous (CaO 4.2 wt%). Moreover, all the samples contain high SiO2 (48 - 60 wt%) and low MgO (0.2 – 0.4 wt%) contents; nevertheless, some differences in these and other elements among the samples can be clearly observed.

17In order to obtain a first insight into the variability of the data set, the Compositional Variation Matrix (CVM) was calculated taking into account the 25 chemical elements (Buxeda, 1999; Buxeda and Kilikoglou, 2003). The CVM provides the total variation, which corresponds to a very high value (tv = 3.55), representative of a polygenic assemblage. The CVM also provides information on the variation that each element introduces into the data set; in this case, the highest values are related to K2O, Rb, and P2O5. The examination of Figure 5b reveals that the variability in K2O is mainly due to the higher values presented by samples DOL4 (4 wt%) and DOL8 (1 wt%) with respect to the rest of the dolia (0.2 – 0.6 wt%). The Rb content is also especially high in DOL4 (100 ppm) and also in DOL3 (36 ppm) and DOL8 (25 ppm), while in the rest of the samples the Rb values oscillate between 6 and 15 ppm. The high variability observed in P2O5 is mainly due to differences in the composition of this element among the dolia analysed. In most of the samples, the P2O5 values range from 0.03 wt% (in DOL3) to 0.2 wt% (in DOL8 and DOL10); however, the values are higher in the specimens DOL5 (0.5 wt%) and DOL3 (0.3 wt%). The variability in P2O5 could be related to post-depositional alterations occurred in the wares presenting higher values. These perturbations are common in pottery deposited in anthropised environments such as cultivated areas or in contact with burials (Maritan and Mazzoli, 2004; Maritan et al., 2009).

18Further multivariate statistical treatment was carried out in order to visualise the distribution of the pottery sherds according to the (dis)similarity of the compositional data. First, the subcomposition Fe2O3 (as total Fe), Al2O3, MnO, TiO2, MgO, CaO, Na2O, K2O, Rb, Th, Nb, Zr, Y, Ce, Ga, V, Zn, Cu, Ni and Cr was log-ratio transformed using SiO2 as divisor (Buxeda, 1999). Then, a Cluster Analysis (CA) was performed by means of the centroid agglomerative method and the squared Euclidean distance. The dendrogram (Fig. 5c) shows that most of the dolia are clustered in a single group, while three individuals (DOL3, DOL4 and DOL8) appear more distanced due to some differences in their chemical composition. The main group includes all the specimens characterised to present high CaO (12 – 19 wt%) and Cr (114 – 171 ppm) contents, and lower Na2O (0.05 – 0.15 wt%), Rb (6 – 15 ppm) and Cu (1 – 7 ppm) values than the rest of the samples. Among these individuals, the composition of DOL5 slightly differs. It consists of a border calcareous fabric (CaO 4.2 wt%) with higher concentrations in Fe2O3 (6 wt%), Al2O3 (18.5 wt%), TiO2 (1.18 wt%), SiO2 (59.7 wt%), Nb (22 ppm), Zr (327 ppm), Ga (22 ppm) and V (109 ppm). DOL3 and DOL4 exhibit low Fe2O3 and Cr content, high CaO, Al2O3, Ga and V values, as well as the highest MnO, Rb, Pb and Zn concentrations. However, main chemical difference between these two dolia concerns the K2O content, which is higher in DOL4 (4 wt%). DOL8 is also calcareous and exhibits low Fe2O3, Al2O3, TiO2, Th, Nb, Ga, V and Ni concentrations. It mainly differs due to the Na2O content (0.34 wt%), which is the highest value of the dataset.

Figure 5 a: Normalised chemical results of the dolia obtained by WD-XRF. a./ Group 1 and 1A averages / Résultats chimiques normalises des dolia obtenus par WD-XRF. a./ Moyenne des groupes 1 et 1A

Figure 5 a: Normalised chemical results of the dolia obtained by WD-XRF. a./ Group 1 and 1A averages / Résultats chimiques normalises des dolia obtenus par WD-XRF. a./ Moyenne des groupes 1 et 1A

Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.

Figure 5 b: Normalised chemical results of the dolia obtained by WD-XRF. b./ Individual samples results / Résultats chimiques normalises des dolia obtenus par WD-XRF. b/. Résultats par échantillons

Figure 5 b: Normalised chemical results of the dolia obtained by WD-XRF. b./ Individual samples results / Résultats chimiques normalises des dolia obtenus par WD-XRF. b/. Résultats par échantillons

Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.

Figure 5 c: Normalised chemical results of the dolia obtained by WD-XRF. c./ Dendrogram resulting from the cluster analysis of the twelve analysed dolia / Résultats chimiques normalises des dolia obtenus par WD-XRF. c./ Dendrogramme résultant de l’analyse de grappe des douze dolia analysés

Figure 5 c: Normalised chemical results of the dolia obtained by WD-XRF. c./ Dendrogram resulting from the cluster analysis of the twelve analysed dolia / Résultats chimiques normalises des dolia obtenus par WD-XRF. c./ Dendrogramme résultant de l’analyse de grappe des douze dolia analysés

Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.

19According to the chemical data, nine of the twelve dolia analysed could be attributed to a single pottery production, which includes dolia found in different sites located in the Drôme, Hérault and Aude departments. The dolia DOL3, DOL4 and DOL8 exhibit major chemical differences and should be considered as different productions, although they were manufactured using similar raw materials to those of the main group, which were processed in an analogous way.

20Even if the variability observed among the dolia analysed is very high, we have to highlight that they consist in very coarse fabrics, which generally exhibit higher intrinsic variability than the fine pastes. Therefore, the compositional differences highlighted by geochemical analysescould be interpreted as both, the use of distinct clay deposits for a single production centre during the period of activity, or the proof of the existence of several workshops located in a neighbouring zone or in nearby areas with similar geological characteristics.

Mineralogical characterization

21The XRD analysis of the twelve dolia has enabled the identification of the mineralogical composition and to obtain an estimation of the equivalent firing temperatures (EFT) (Cultrone et al., 2001; Maggetti et al., 2011). The diffractograms point to the presence of quartz, calcite and illite-muscovite as the main primary mineralogical phases. K-feldspars and plagioclase appear in a minor degree. Hematite is the iron oxide found in the dolia fired in oxidising conditions, while magnetite was found on DOL3, the single specimen presenting a dark matrix derived of its firing in reducing conditions (Fig. 6). The only clear firing phase identified is gehlenite, which has been detected in all the samples except in the border-calcareous dolium DOL5.

Figure 6: Diffractograms of representative dolia / Diffractogramme représentatif des dolia étudiés

Figure 6: Diffractograms of representative dolia / Diffractogramme représentatif des dolia étudiés

cal= calcita; gh= gehlenite; hm= hematite; ill= illite- muscovite; kfs= K-feldspar; pl= plagioclase; mag: magnetite; qtz= quartz. / cal= calcita; gh= gehlenite; hm= hematite; ill= illite- muscovite; kfs= K-feldspar; pl= plagioclase; mag: magnetite; qtz= quartz.

22Gehlenite is a metastable phase produced in calcareous pastes fired at temperatures up to 800ºC. The crystallisation of gehlenite derives of the thermal reactions that occur in the primary Ca-rich material until it becomes a fired clay ceramic (Cultrone et al., 2001; Trindade et al., 2009; Maggetti et al., 2011; Fabbri et al., 2014). Under prolonged firings, carbonates decompose through decarbonation at temperatures between 600ºC and 800°C, depending on the type and grain-size. During this process, carbon dioxide gas (CO2) is released, and free-lime (CaO) is formed. Up to 800ºC the dehydroxylation of the clay minerals to form the pseudo-amorphous phase of meta-clay takes place. During this process, the free-lime of the decarbonated calcite reacts with Si and generates nucleation and crystallization of several metastable phases such as gehlenite and diopside, which continuously react with increasing temperature. Gehlenite starts to crystallise at temperatures ranging from 800ºC to 850ºC, while diopside occurs at c. 800-850ºC but in very low concentration and forms drastically at 900-950ºC. According to these data, the equivalent firing temperature of the twelve dolia analysed has been estimated around 850ºC. A lower firing temperature (below 850ºC) is suggested to the dolium DOL5 since its diffractogram only exhibits primary phases.

5. Discussion: Characteristics of an unusual production

Signs of technological optimization

23The study of the technological characteristics of dolia with calcite tempers points to the unprecedented technical optimization of doliaire craftsmanship, highlighting once again the extraordinary sophistication of this production. In particular, this progress is perceptible in the “recipe” used for paste preparation. The respective quantities of clay and temper were determined based on the comparative tables generally used for sedimentary petrographic descriptions (Terry and Chilingar, 1955). According to these tables, the proportion of calcite temper is around 20 %. In this way, we estimated that nearly 560 kg of unfired clay and 160 kg of temper were required for the manufacture of a dolium of 15 hl (Carrato, 2017, p.133).

24The abundance of calcite deposits may explain the predominant use of this temper, but from a technical perspective, it nonetheless raises difficulties during firing, as primary calcite starts to decompose at 700-750°C and their total decomposition creates porosities. In addition, clay and calcite present divergent reactions during firing, putting pressure on the cohesion of the paste; clay reduces in volume when heated whereas calcite dilates (Cultrone, et al., 2001; Trindade, et al., 2009, 2010; Maggetti, et al., 2011; Fabbri, et al., 2014). XRD analyses were undertaken to rule out perturbations in the chemical and mineralogical composition of the samples due to weathering processes, but they also allow for the detection of primary crystalline phases, or modifications linked to firing. All the dolia analysed show predominant primary crystalline phases and only gehlenite as incipient firing phase, suggesting low firing temperatures in all the cases. However, thin section observation shows that the crystals near the surface, in direct contact with the laboratory atmosphere, were often destabilized, whereas the elements near the heart of the paste did not undergo any modification. These differences are due to the fact that the dolia have very thick walls and the effects of the temperature during the firing were smaller in the core of the walls than in the surfaces.

25The low temperatures required to not destabilize the calcareous inclusions involved a great control of the firing process. At the same time, this low temperature undoubtedly represented significant fuel saving, indicating the will of the workshop owners to reduce production costs by implementing technological improvements. On the other hand, this decrease in temperature did not guarantee optimal clay firing, although it seems to improve the fracture toughness, which was also enhanced by adding large amounts of non-plastic inclusions of different size and composition to a Fe-rich clayey material (Kilikoglou, et al., 1995, 1998; Vekinis and Kilikoglou, 1998; Tite, et al., 2001). The study of the composition of the group 1 paste suggests that the choice of calcite may be interpreted as a deliberate act aiming to reduce the quantity of fuel necessary for dolia firing and consequently, to reduce production costs.

An ideal geographic implantation; the hypothesis of the Middle Rhône Valley

26The calcite temper group 1 presents a more extensive diffusion zone than all of the other pastes identified in Narbonese Gaul (Fig. 7). It is recorded from Lyon to Narbonne, in the Rhône Valley and along the Languedoc coastline, over nearly 400 km. Specimens from hinterland sites, such as the Middle Hérault Valley, have also been identified. Lastly, a specimen was found near Marseille, on the site of the Garanne villa at Berre-l’Étang (Bouches-du-Rhône), which seems to correspond to the easternmost evidence of this paste.

Figure 7: Diffusion map of group 1 dolia with calcite temper and hypothetical location of the production workshop (H. Bohbot and C. Carrato) / Carte de diffusion de dolia du groupe 1 à dégraissant de calcite et emplacements supposés de la zone de production (H. Bohbot et C. Carrato)

Figure 7: Diffusion map of group 1 dolia with calcite temper and hypothetical location of the production workshop (H. Bohbot and C. Carrato) / Carte de diffusion de dolia du groupe 1 à dégraissant de calcite et emplacements supposés de la zone de production (H. Bohbot et C. Carrato)
  • 5 For example, we refer here in particular to about fifteen alleged workshops in France with elements (...)

27The size of these dolia production installation(s) must have been considerable, in view of the diffusion area of dolia and the number of vessels. Paradoxically, archaeological research has not revealed any such installations in Narbonese Gaul. In fact, the identification and excavation of dolia workshops are almost inexistent in the province. As mentioned above, only the Saint-Bézard workshop at Aspiran (Mauné, et al., 2006) has yielded two large kilns, and traces of a third kiln, clearly designed to fire these large terracotta containers. Other workshops are often assumed to have existed, but no comparable production structures have ever been brought to light5.

28Although the workshop cannot be accurately located, the available markers nonetheless allow us to suggest a working hypothesis about the production area location. By combining data related to the location of the archaeological evidence and the petrographic composition of the samples analysed, we can assume that the production zone is near the Rhône, where calcareous formations and karstic caves are abundant. In addition, the composition of the alluvial sand, with few elements from the metamorphic and granitic bedrock of the Massif Central, could point to the hypothesis that the workshop or workshops should be probably located on the left bank of the Rhône. However, metamorphic and granitic elements of this type are so rare that a production zone on the Rhône could be ruled out. These markers may imply that the production zone is on a tributary of the Rhône, with a left bank confluence.

29The two most likely production zones going up the Rhône from Arles to Lyon are the Rhône-Drôme confluence and Rhône-Roubion. These two sectors are about fifteen kilometres apart, and contain alluvial sands with this composition, as well as large quantities of karstic calcite, in particular in the caves of the Diois Valley, near the Drôme River. Moreover, the geological map reveals that Pliocene clay deposits are very abundant in this sector, mainly on the right bank of the Drôme between Livron-sur-Drôme and Crest and on both banks of the Roubion between Saint-Marcel-lès-Sauzet and La Laupie (Fig. 8).

Figure 8: Geological map of the two production areas suggested by the cross of archaeological, archaeometrical, geological and toponymic data (BRGM, 2019) / Carte géologique des deux zones de production suggérées par le croissement des données archéologiques, archéométriques, géologiques et toponymiques (BRGM, 2019)

Figure 8: Geological map of the two production areas suggested by the cross of archaeological, archaeometrical, geological and toponymic data (BRGM, 2019) / Carte géologique des deux zones de production suggérées par le croissement des données archéologiques, archéométriques, géologiques et toponymiques (BRGM, 2019)
  • 6  Mirmande, Hauterives, Saillans, Crest, le Bourg-lès-Valence, Loriol, Montboucher, Chabeuil, Érôme, (...)
  • 7  Outside this area, the geological map of Montélimar from 1979 also mentions the exploitation of cl (...)

30The toponymy in the northern sector of the Drôme-Rhône confluence indicates a brick and tile tradition in this sector. Three hamlets called “La tuilière – tilemaking”, “La tuilerie – the tilery” or “La potière-potter” and seven modern tileries were identified in this zone, which is about fifteen kilometres long. In addition, the location of these sites on the edge of the Drôme and the abundance of raw materials make this sector an ideal choice for manufacturing. A study of the archives and early geological maps confirms that the history of pottery activities in this zone goes back a long way. In a book dating to 1835, M. Delacroix wrote that the production of bricks and tiles was so important that it supplied the whole department with construction materials (Delacroix, 1835, p. 374). In the same passage he cites the townlands with the most active production6. We observe that the latter are situated along the left bank of the Rhône or at the confluence of the Rhône and the Drôme. The geological map of Crest dated to 1977 also states that the numerous tileries used the Pliocene clay between Livron-sur-Drôme and Crest7.

31In the southern sector of the Rhône-Roubion confluence, the former Saint-Marcel-lès-Sauzet priory was named in loco Fellinis in 985 by Count Lambert, which is the modern trend of Figlinae latin name for potter workshop. The change of name came about during the next generation when Sauzet castle became a county seat (Bois, 1995). In a document dated to 1789, significant “tile earth” deposits are declared to be unused due to a lack of wood. The Roubion was floatable like the Drôme.

32In our opinion, all of these elements reinforce the possibility that raw material deposits were exploited in ancient times in the middle Rhône Valley, between Montélimar and Loriol-sur-Drôme. The position of these deposits near a river would also have facilitated production exportation by river. Unfortunately, the archaeological map (Planchon, et al., 2010) reveals little information about ancient settlement in this sector. The rare operations carried out there have mainly recorded an important sedimentary cover of the alluvial plain, which seems however to be less important in the southern production zone. However, if one or several workshops using this paste with calcite temper were in operation during ancient times, then traces of this pre-industry may not have totally disappeared in this zone.

33In order to validate these assumptions, magnetic prospections are planned in the coming years. However, this research should be preceded by a mapping work (old and recent maps) to limit the size area of intervention. This will include the removal of urbanised areas or areas that have undergone major redevelopment since Antiquity but also the target of the most suitable areas taking into account topographical data. Likewise, if the area to be explored remains too large, prospecting may be limited to a 10-20 km strip along the course of the rivers, considering the importance of proximity to the river tracks for this activity.

A providential chronology; a concomitant development of provincial viticulture

34The comparison of the chronology of the advent of this group of paste with the development of wine making and olive growing in the Narbonese province points to a similar expansion. The evidence seems to imply that the centre or centres of dolia production with calcite temper developed in response to the significant increase in wine and oil production during the Tiberian period, between 20 and 40 A.D.

  • 8 In the Drôme, storehouses of dolia with this paste, located near the alleged production zone, are d (...)

35The scale of manufacture characteristic of this production area is perfectly clear when we examine the general chronology of doliaire manufacture during the Roman period in Narbonese Gaul. According to our knowledge, it appears that the earliest identified workshops, dating from the end of the first century B.C. and the first decades of the first century A.D., seem to correspond to domanial production structures with micro-regional and regional diffusion (Carrato, 2017, pp .159-164). The emergence of a workshop during the period ranging from 20 to 40 A.D.8, with much more intensive production activity than previously recorded, thus upsets this relatively stable pattern. This stage seems to represent a rupture in the production process, although it was not possible to determine whether small workshops regressed as a result of this new competition.

36In any event, the study of the sites from this period shows that the use of dolia with calcite temper became widespread in most of the Narbonese province, apart from the southeast quarter. From the Claudian period onwards, these workshops became the sole suppliers of the dolia used in the whole Rhône Valley, and also became predominant in the western zone of the province, as far as Narbonne.

37As an example of this supremacy, the 350 dolia from storehouse 1 from the Vareilles villa in Paulhan (Hérault), one of the largest known storehouses in Narbonese Gaul (Mauné, 2003a), come from this workshop, which is theoretically located nearly 250 km from there.

38The acceleration of dolia production as a result of the creation of large production centres with wide diffusion zones corresponds chronologically to the expansion of viticulture, and to a lesser extent of olive growing (Brun, 2005 ; Brun and Laubenheimer, 2001). This point is perfectly illustrated by the increase in storage capacities from the first half of the first century B.C. onwards, closely followed by the multiplication of the number of operations (Carrato, to be published) and by the decrease in importations of Spanish wine amphorae at the end of the Augustan period (Duperron, 2014, p. 140, pp. 145-146) and by the emergence at the same time of workshops producing wine amphorae (Laubenheimer, 1985 ; Laubenheimer and Schmitt, 2009 ; Mauné, 2009 ; 2013). From the second half of the second century A.D., the number of newly created centres decreased sharply, as did storage capacities and the number of amphora workshops, clearly demonstrating the early stages of the third century economic downturn.

6. Conclusions: the proof of pre-industrial dolia manufacture

39This overview enabled us to show that the archaeometric analysis of a considerable batch of dolia fragments from use contexts can lead to the extraction of unprecedented markers of the production and transport conditions of the largest known ceramic vessel during antiquity.

40Through the study of significant corpus of dolia with calcite temper, we can decipher the means used by the economic actors of the Narbonese province to meet the exponential demand for these containers, alongside the unprecedented development of local viticulture from the first quarter of the first century A.D. onwards. The development of the alleged large workshop supports the intensity of production, made possible by certain technical optimizations leading to the reduction of firing time, and the scale of the diffusion range of the dolia, probably favoured by an implantation along the Rhône in order to take advantage of this major circulation route.

41On account of the degree of optimization of production means and the constant quest for profitability, this unprecedented discovery revives the debate on the characterization of the Roman economy, which periodically presents pre-industrial characteristics. This overview also underscores once again the interest of studying objects on a regional scale, in order to bring to light poorly known economic mechanisms, which remain inaccessible for global studies.

Haut de page

Bibliographie

Bois, M., 1995. Le prieuré de Saint-Marcel-lès-Sauzet (Drôme), In: Congrès Archéologique de France 150e session Moyenne vallée du Rhône 1991. p. 231-238.

Bonnet, C. and Horry, A., 2010. Céramiques de la basse vallée de la Drôme du ier s. av. J.-C. au vie s. apr. J.-C. ; les sites de Saint-Martin et l’Hortal à Chabrillan (Drôme). Revue Archéologique de la Narbonnaise, 43, p. 255-290.

Brun, J.-P., with coll. of Borréani, M., 1999. Le Var. Paris, CNRS.

Brun, J.-P., 2005. Archéologie du vin et de l’huile en Gaule romaine. Paris, Errance.

Brun, J.-P., and Laubenheimer, F., eds., 2001. La viticulture en Gaule. Paris, CNRS.

Buxeda i Garrigós, J., 1999. Alteration and contamination of archaeological ceramics: the perturbation problem. Journal of Archaeological Sciences, 26, p. 295-313.

Buxeda i Garrigós, J. and Kilikoglou, V., 2003. Total Variation as a Measure of Variability in Chemical Data Sets, a L. van Zelts (ed.), Patterns and Process: a Festschrift in honour of Dr. Edward V. Sayre, Smithsonian Centre for Materials Research and Education, Suitland, Maryland, pp. 185-198.

Cambon, C., Carozza, L., Malet, L., Provost, M., Devillers, B., Gangloff, N., Ruffat, A. and H. and Salinier, J.-F., 1995. Le Tarn. Paris, CNRS.

Carrato, C., 2017. Le dolium en Gaule narbonnaise (ier a.C.-iiie p.C.). Contribution à l’histoire socio-économique de la Méditerranée nord-occidentale. Bordeaux, Ausonius Edition.

Carrato, C., to be published. Dynamique de la production vinicole en Gaule Narbonnaise à l’époque romaine à partir de l’étude des capacités de stockage (ier s. av.-ve s. apr. J.-C.). ». In Report on the XIXth International Congress of Classical Archaeology, Cologne/Bonn, 22-26 may 2018.

Checa, M. U., 1901. Des bouteilles de seize milles litres. Lectures pour tous, Revue Universelle illustrée, 3e année n°2 novembre 1901, p. 149-154.

Cultrone, G., Rodriguez-Navarro, C., Sebastian, E., Cazalla, O., and De la Torre, M.J., 2001. Carbonate and silicate phase reactions during ceramic firing. European Journal of Mineralogy, 13, p. 621-634.

Delacroix, N., 1835. Statistique du département de la Drome. Valence: de Borel.

Dellong, E., with coll. of Moulis, D. and Farré, J., 2003. Narbonne et le Narbonnais, Paris, CNRS.

Descemet, M. C., 1880. Inscriptions doliaires latines ; marques de briques relatives à une partie de la Gens Domitia. Paris, E. Thorin.

Duperron, G., 2014. Arles et Lyon, ports fluviaux de l’Empire romain. Le commerce sur l’axe rhodanien du Ier s. av. J.-C. au VIIe s. apr. J.-C., Vol.1 synthèse. Ph. D. Université Montpellier Paul Valéry.

Fabbri, B., Gualtieri, S., and Shoval, S. 2014. Review. The presence of calcite in archeological ceramics. Journal of the European Ceramic Society, 34, p. 1899-1911.

Gasperoni, T., 2003. Le fornaci dei Domitii; ricerche topografiche a Mugnano in Teverina. Viterbo, Universita degli studi della Tuscia.

Gasperoni, T., 2005. Nuove acquisizioni dai praedia dei Domitii nella valle del Fosso del Rio. In: C. Bruun and F. Chausson, eds. 2005. Interpretare i bolli laterizi di Roma e della Valle del Tevere; produzione, storia economica e topografia: Atti del convegno. Roma 2000. Roma: Acta Instituti Romani Finlandiae. p. 103-120.

Giannopoulou, M., 2010. Pithoi; technology and history of storage vessels through the ages. Oxford: Archaeopress.

Gruat, P., Malige, G. and Vidal, M., with coll. of ASPAA, 2011. L’Aveyron. Paris, CNRS.

Guillaume, M., 2004. La villa de Saint-Ariès. In: A. Jourdin-Annequin, dir. 2004. Atlas culturel des Alpes occidentales ; De la préhistoire à la fin du Moyen Âge. Paris, Picard, p. 159-161.

Kilikoglou, V., Vekinis, G., and Maniatis, Y., 1995. Toughening of ceramics earthenwares by quartz inclusions: an ancient art revised, Acta Metallurgica et Materialia, 43, p. 2959-2965.

Kilikoglou, V., Vekinis, G., Maniatis, Y., and Day, P., 1998. Mechanical performance of quartz-tempered ceramics: Part I, Strength and toughness. Archaeometry, 40, p. 261-277.

Kotarba, J., Castellvi, G. and Mazière, F., 2007. Les Pyrénées-Orientales. Paris, CNRS.

Laubenheimer, F., 1985. La production des amphores en Gaule narbonnaise. Paris, Les Belles Lettres.

Laubenheimer, F. and Schmitt, A., 2009. Amphores vinaires de Narbonnaise ; production et grand commerce. Création d’une base de données géochimiques des ateliers. Lyon, Maison de l’Orient et de la Méditerranée Jean Pouilloux.

Lazzeretti, A., and Pallecchi, S., 2005. Le figlinae polivalenti; la produzione di dolia e di mortaria bollati. In: C. Bruun and F. Chausson, eds. 2005. Interpretare i bolli laterizi di Roma e della Valle del Tevere ; produzione, storia economica e topografia: Atti del convegno. Roma 2000. Roma: Acta Instituti Romani Finlandiae. p. 213-227.

Maggetti, M., Neururer, C., Ramseyer, D., 2011, Temperature evolution inside a pot during experimental surface (bonfire) firing. Applied Clay Science, 53, p. 500-508.

Malignas, A., Sachot, G., Sanchez, C. and Convertini, F., 2008. Découverte d’un atelier de céramiques culinaires à Laure-Minervois (Aude). In: L. Rivet, ed. 2008. Actes du congrès de la SFECAG. Escala-Empuries 2008. Marseille: SFECAG. p. 803-811.

Maritan, L., and Mazzoli, C., 2004, Phosphates in archaeological finds: implications for environmental conditions of burial. Archaeometry, 46, p. 673-683.

Maritan, L., Angelini, I., Artioli, G., Mazzoli, C., and Saracino, M., 2009, Secondary phosphates in the ceramic materials from Frattesina (Rovigo, North-Eastern Italy). Journal of Cultural Heritage, 10, p. 144-151.

Martínez Ferreras, V., Capelli, C., Cabella, R., and Nieto Prieto, X., 2013. From Hispania Tarraconensis (NE Spain) to Gallia Narbonensis (S France). New data on Pascual 1 amphora trade in the Augustan period. Applied Clay Science, 82, p. 70-78.

Martínez Ferreras, V., 2017. Analyses géochimiques des pâtes de dolia d’époque romaine provenant de Gaule Narbonnaise. In: C. Carrato. 2017. Le dolium en Gaule narbonnaise (Ier a.C.-IIIe p.C.). Contribution à l’histoire socio-économique de la Méditerranée nord-occidentale. Bordeaux: Ausonius Edition. p. 22-241.

Mauné, S., 2003a. La villa gallo-romaine de Vareilles à Paulhan (Hérault ; fouille de l’Autoroute A75) un centre domanial du Haut-Empire spécialisé dans la viticulture ?. In: S. Lepetz and V. Matterne, eds. 2003. Cultivateurs, éleveurs et artisans dans les campagnes de Gaule romaine, matières premières et produits transformés: Actes du Ve colloque AGER, Compiègne 2002. Amiens, Revue Archéologique de Picardie. p. 309-337.

Martínez Ferreras, V., 2003b. Paysage et usages du sol dans la région de Béziers (Gaule narbonnaise) entre le ier s. av. J.-C. et le iiie s. apr. J.-C. In: L. Labruna and R. Pierobon, dir. 2003. Paesaggio e utilizzazione della terra: diritto, economia, società: Atti del convegno internazionale, Roma 1999. Roma: Quaderni camerti di studi romanistici. p. 250-294.

Martínez Ferreras, V., 2009. Recherches récentes sur les ateliers de potiers de Gaule narbonnaise ; contribution à l’histoire socio-économique d’une province romaine (ier s. av. J.-C.-iiie s. apr. J.-C.). Habilitation à diriger des recherches, Montpellier-Lattes.

Martínez Ferreras, V., 2014. Entre Thongue et Libron (Hérault), une forêt aux portes de la colonie romaine de Béziers (Ier-IIIe s. apr. J.-C.). In: F. Favory, dir. 2014. Silva et Saltus en Gaule romaine: dynamique et gestion des forêts et des zones rurales marginales: Actes du VIe colloque AGER, Rennes 2004. Besançon, Presses Universitaires de Franche-Comté. p. 147-174.

Mauné, S., Bourgaut, R., Lescure, J., Carrato, C. and Santran, C., 2006. Nouvelles données sur les productions céramiques de l’atelier de Dourbie à Aspiran (Hérault) (première moitié du Ier siècle apr. J.-C.). In: L. Rivet, ed. 2006. Actes du congrès de la SFECAG, Pézenas 2006. Marseille, SFECAG. p.157-188.

Odiot, T., 1996. Donzère-Le Molard. In: Harur, Formes de l’habitat rural en Gaule narbonnaise 3, Spécial villa romaine. Juans-les-Pins, APDCA. p. 38-62.

Pagès, G., 2008. La métallurgie du fer en France méditerranéenne de l’Antiquité au début du Moyen Âge: jalons d’une approche interdisciplinaire. Ph. D. Université Montpellier III Paul Valéry.

Pallecchi, S., 2012. Appunti sull’organizzazione del lavoro negli stabilimenti romani per la produzione di ceramica pesante. In: A. Ciacci, P. Rendini and A. Ziffereo, eds. 2012. Archeologia della vite e del vino in Toscana e nel Lazio: dalle tecniche dell’indagine archeologica alle prospettive della biologia molecolare. Borgo San Lorenzo, All’insegna del gilio. p. 469-484.

Passelac, M., 2007. Ateliers céramiques domaniaux du bassin audois au Haut-Empire. In: J. P. Nolla Brufau, dir. 2007. Pottery workshops and agricultural productions. Girona, Universitat de Girona, p. 115-137.

Planchon, J., Bois, M. and Conjard-Réthoré, P., 2010. La Drôme. Paris, CNRS.

Romero Vidal, A. and Cabasa Calpe, S., 1999. La Tinajería tradicional en la cerámica española. Barcelona: Ediciones Ceac.

Rothé, M.-P. and Tréziny, H., 2005. Marseille et ses alentours. Paris, CNRS.

Steinby, E.-M., 1993. Ricerche sull’industria doliare nelle aree di Roma e di Pompei; un possibile modello interpretativo ?. In: C. Zaccaria, ed. 1993. I Laterizi di età romana nell’area nordadriatica: Atti del convegno, Udine 1987. Roma: l’Erma di Bretschneider, p. 9-14.

Tallah, L., 2004. Le Luberon et Pays d’Apt. Paris, CNRS.

Terry, R. D. and Chilingar, G. V., 1955. Summary of Concerning some additional aids in studying sedimentary formations. Journal of Sedimentary Petrology, 25, pp. 229-234.

Tite, M.S., Kilikoglou, V. and Vekinis, G., 2001. Review Article. Strength, toughness and thermal shock resistance of ancient ceramics, and their influence on technological choice. Archaeometry, 43, 3, p. 301-324.

Tremoleda, J., Castanyere, P., Simon, J., Andrea Ferrer, A. 2018. La Bòbila d’Ermedàs (Cornellà del Terri). Una indústria de producció d’època romana al Pla de l’Estany, Girona, Universitat de Girona, Institut de Recerca Històrica.

Trindade, M.J., Dias, M.I., Coroado, J., and Rocha, F. 2009. Mineralogical transformations of calcareous rich clays with firing: A comparative study between calcite and dolomite rich clays from Algarve, Portugal. Applied Clay Science, 42, p. 345-355.

Ugolini, D. and Olive, C., with coll. of Gomez, E., 2012. Béziers Paris: CNRS.

Vaz Pinto, I., 1997. Dolia de São Cucufate et jarres modernes de l’Alentejo ; essai d’ethnoarchéologie. In: R. Etienne and F. Mayet, eds. 1997. Itinéraires Lusitaniens ; trente années de collaboration luso-françaises: Actes de la réunion, Bordeaux 1995. Paris, de Boccard. p. 111-156.

Vekinis, G. and Kilikoglou, V., 1998. Mechanical performance of quartz-tempered ceramics: Part II, hertzian strength, wear resistance and applications to ancient ceramics. Archaeometry, 40, 2, p. 281-292.

Haut de page

Notes

1 Known as In dolia fugularem artem discere by Plato, Gorgias, 514e and Lachès, 187b, or Erasmus, Les Adages, 515.1.6.15.

2 See in particular Descemet, 1880 ; Steinby, 1993 ; Gasperoni, 2003 ; 2005 ; Lazzeretti and Pallecchi, 2005; Pallecchi, 2012; Carrato, 2017, pp. 164-174.

3 DOL1, DOL2 : La Nautique, Narbonne (Aude) ; DOL3, DOL5 : La Barque, Allan (Drôme) ; DOL4, DOL7 : Le Molard, Donzère (Drôme) ; DOL6 : La Bâtie-Rolland (Drôme) ; DOL8 : L’Auribelle basse, Pézenas (Hérault) ; DOL9 : Saint-Sauveur, Lattes (Hérault) ; DOL10 : Le Pendut, Magalas (Hérault) ; DOL11 : Saint-Bézard, Aspiran (Hérault) ; DOL12 : Vareiles, Paulhan (Hérault).

4 The inclusions in the fine fraction correspond to the smallest grains (< 0.125 mm) and are abundant in the clay matrix. Conversely, the coarse fraction corresponds to the largest inclusions (> 0.125 mm), which may appear as natural components included in the clayey raw material procured for the manufacture, or deliberately added and mixed with the clayey raw material by the potters.

5 For example, we refer here in particular to about fifteen alleged workshops in France with elements discovered during prospection or excavations. The identification of some of them is based on solid markers (stamps, traces of large kilns, paste with local temper, large basin for clay preparation), such as: le Petit-Clos II-III in Perpignan (Pyrénées-Orientales) (Kotarba, et al., 2007, pp. 482-489); Salauze in Laure-Minervois (Aude) (Malignas, et al., 2008); Mas de Bourgade in Servian (Hérault) (Mauné, 2003b) and Capitou in Servian (Hérault) (Mauné, 2014, p. 157). Other identifications are based on less reliable evidence, in particular on overfired dolia elements, which were frequently reused. This is the case for the sites of: Poujols in Laissac (Aveyron) (Gruat, et al., 2011, p.187); rue de la Piale in Albi (Tarn) (Cambon, et al., 1995, p. 59); Fount-de-Rome in Fleury-Saint-Pierre (Aude) (Dellong, 2003, pp. 537-539); Téoulet in Villepsy (Aude) (Passelac, 2007, p.115); Clairac I in Béziers (Hérault) (Ugolini and Olive, 2012); Grangettes in Puissalicon (Hérault) (info. G. Fédière); La Garcine in Mirabeau (Vaucluse) (Tallah, 2004, p. 279); Tournon in La Bouilladisse (Bouches-du-Rhône) (Rothé and Trezigny, 2005, p. 789); la Pinède in Castellet (Var) (Brun, 1999, pp. 314-416) and Saint-Ariès in Ventavon (Hautes-Alpes) (Guillaume, 2004).

6  Mirmande, Hauterives, Saillans, Crest, le Bourg-lès-Valence, Loriol, Montboucher, Chabeuil, Érôme, Ponsas, Eymeux, la Garde-Arhémar and Savasse.

7  Outside this area, the geological map of Montélimar from 1979 also mentions the exploitation of clay, but also marls from the Upper Aptian, the Upper Cretaceous and the Pliocene, on the banks of the Rhône by brickworks or pottery centres. On the Dieulefit map dating from 1969, white or grey clay levels made up of 90% kaolinite were used for making fine pottery. On the other hand, for the rest of the Drôme department, the geological maps of Nyons, Valréas, Die, Luc-en-Diois and Valence do not mention any exploited clay deposits.

8 In the Drôme, storehouses of dolia with this paste, located near the alleged production zone, are dated to the second quarter or the middle of the first century A.D. at the earliest (in Chabrillan, Saint-Martin 1 (Bonnet and Horry, 2010) and Donzère, Le Molard (Odiot, 1996). In the Hérault, this paste is evidenced in dolia in wine storehouses installed during the same period. At the site of Saint-Bézard in Aspiran (Carrato, 2017) for example, they are not used in the first phase of the storehouse dated to 17/25 A.D. – where locally made dolia are used – but are massively used during the reconstruction of the southern angle of the storehouse at around 50/60 A.D. In the neighbouring villa of Vareilles at Paulhan (Mauné, 2003a), the large storehouse 1 equipped with dolia with this paste was built during the fifth decade A.D. Further away, at the site of Port la Nautique in the Aude, dolia from this production appear in levels dated to 20-40 A.D. (Carrato, 2017, p. 161).

Haut de page

Table des illustrations

Titre Figure 1: Dolium from the Bernardins villa in Montélier (Drôme) (capacity: 2,463 l) conserved in the Museum of art and archaeology of Valence / Dolium provenant de la villa des Bernardins à Montélier (Drôme) (capacité : 2 463 l) conservé au musée d’art et d’archéologie de Valence
Crédits Cl. C. Carrato.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-1.jpg
Fichier image/jpeg, 603k
Titre Figure 2: Distribution map of the sites in the studied corpus (H. Bohbot and C. Carrato) / Carte de répartition des sites du corpus d’étude (H. Bohbot et C. Carrato)
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-2.jpg
Fichier image/jpeg, 1,9M
Titre Figure 3: Macroscopic views of fragments of group 1 paste (1,5x) / Vues macroscopiques de fragments de pâte du groupe (1,5x)
Légende The numbers refer to the sites listed in Fig. 2. / Les numéros renvoient aux sites répertoriés dans la Fig. 2.
Crédits Cl. C. Carrato.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-3.jpg
Fichier image/jpeg, 908k
Titre Figure 4: Microphotographs of thin sections taken under plane polars (PPL) and crossed polars (XPL) at the same magnification (20x); a) representative individuals of the main fabric identified; b) the dolium DOL8; c) the dolium DOL3 / Microphotographies des lames minces en lumière polarisante non analysée (LPNA) et en lumière polarisante analysée (LPA) au même grossissement (20x) ; a. Individus représentatifs de la même fabrique ; b. le dolium DOL 8 ; c) le dolium DOL3 (V. Martínez Ferreras)
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-4.jpg
Fichier image/jpeg, 4,4M
Titre Figure 5 a: Normalised chemical results of the dolia obtained by WD-XRF. a./ Group 1 and 1A averages / Résultats chimiques normalises des dolia obtenus par WD-XRF. a./ Moyenne des groupes 1 et 1A
Légende Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-5.jpg
Fichier image/jpeg, 200k
Titre Figure 5 b: Normalised chemical results of the dolia obtained by WD-XRF. b./ Individual samples results / Résultats chimiques normalises des dolia obtenus par WD-XRF. b/. Résultats par échantillons
Légende Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-6.jpg
Fichier image/jpeg, 413k
Titre Figure 5 c: Normalised chemical results of the dolia obtained by WD-XRF. c./ Dendrogram resulting from the cluster analysis of the twelve analysed dolia / Résultats chimiques normalises des dolia obtenus par WD-XRF. c./ Dendrogramme résultant de l’analyse de grappe des douze dolia analysés
Légende Concentrations of major and minor oxides are in %, other minor and trace elements are in ppm, and the lost on ignition (LOI) in %. / Résultats chimiques normalises des dolia obtenus par WD-XRF. Les concentrations majeure et mineur des oxydes sont exprimées en %, les autres mineurs et éléments traces sont en ppm, et la perte au feu (PAF) est en %.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-7.jpg
Fichier image/jpeg, 118k
Titre Figure 6: Diffractograms of representative dolia / Diffractogramme représentatif des dolia étudiés
Légende cal= calcita; gh= gehlenite; hm= hematite; ill= illite- muscovite; kfs= K-feldspar; pl= plagioclase; mag: magnetite; qtz= quartz. / cal= calcita; gh= gehlenite; hm= hematite; ill= illite- muscovite; kfs= K-feldspar; pl= plagioclase; mag: magnetite; qtz= quartz.
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-8.jpg
Fichier image/jpeg, 486k
Titre Figure 7: Diffusion map of group 1 dolia with calcite temper and hypothetical location of the production workshop (H. Bohbot and C. Carrato) / Carte de diffusion de dolia du groupe 1 à dégraissant de calcite et emplacements supposés de la zone de production (H. Bohbot et C. Carrato)
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-9.jpg
Fichier image/jpeg, 1,9M
Titre Figure 8: Geological map of the two production areas suggested by the cross of archaeological, archaeometrical, geological and toponymic data (BRGM, 2019) / Carte géologique des deux zones de production suggérées par le croissement des données archéologiques, archéométriques, géologiques et toponymiques (BRGM, 2019)
URL http://journals.openedition.org/archeosciences/docannexe/image/6257/img-10.jpg
Fichier image/jpeg, 2,0M
Haut de page

Pour citer cet article

Référence papier

Charlotte Carrato, Verónica Martínez Ferreras, Jean-Marie Dautria et Michèle Bois, « The biggest Opus doliare production in Narbonese Gaul revealed by archaeometry (first to second centuries A.D.) »ArcheoSciences, 43-1 | 2019, 69-82.

Référence électronique

Charlotte Carrato, Verónica Martínez Ferreras, Jean-Marie Dautria et Michèle Bois, « The biggest Opus doliare production in Narbonese Gaul revealed by archaeometry (first to second centuries A.D.) »ArcheoSciences [En ligne], 43-1 | 2019, mis en ligne le 01 janvier 2022, consulté le 29 mars 2024. URL : http://journals.openedition.org/archeosciences/6257 ; DOI : https://doi.org/10.4000/archeosciences.6257

Haut de page

Auteurs

Charlotte Carrato

ATER-Collège de France, associated researcher, UMR5140-ASM /LabEx ARCHIMEDE-MontpellierLecturer-researcher, ERAAUB – University of BarcelonaLecturer emeritus, Geoscience Laboratory – University Montpellier IIPhD in medieval archaeology, Archéo-Drôme Association

Verónica Martínez Ferreras

Lecturer-researcher, ERAAUB – University of BarcelonaLecturer emeritus, Geoscience Laboratory – University Montpellier IIPhD in medieval archaeology, Archéo-Drôme Association

Articles du même auteur

Jean-Marie Dautria

Lecturer emeritus, Geoscience Laboratory – University Montpellier IIPhD in medieval archaeology, Archéo-Drôme Association

Michèle Bois

PhD in medieval archaeology, Archéo-Drôme Association

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search