- Email: [email protected]
Analyses of the white barbotine decoration of two Roman pottery groups
CLAY-02695; No of Pages 5 E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
1
Contents lists available at SciVerse ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Analyses of the white barbotine decoration of two Roman pottery groups Eszter Harsányi a, István Sajó b, György Szakmány c,⁎, Zsolt Bendő c a b c
Archäologisches Landesmuseum Baden-Württemberg, Außenstelle Rastatt, Lützowerstraße 10, 76437 Rastatt, Germany Institute of Materials and Environmental Chemistry, Research Center for Natural Sciences, Hungarian Academy of Sciences, H-1025 Budapest, Pusztaszeri út 59-67, Hungary Department of Petrology and Geochemistry, Eötvös Loránd University, H-1117 Budapest, Pázmány P. sétány 1/c, Hungary
a r t i c l e
i n f o
Article history: Received 26 September 2012 Received in revised form 20 June 2013 Accepted 26 June 2013 Available online xxxx Keywords: White barbotine Roman pottery Black coated ware Samian Ware Trier Rheinzabern
a b s t r a c t White barbotine is a rare decoration form in Roman pottery production. It was applied on three main pottery groups from the 2nd to the 4th century AD: on the Samian ware from Rheinzabern, on the black coated ware from central Gaul and on the motto beakers from Trier. This kind of decoration has not yet been analysed with scientific methods. As a part of a research on the black coated ware in Pannonia some motto beakers from Trier and some samples of Samian ware from Rheinzabern were analysed with XRD, petrographic and EDS-EDX methods to get information about the technique and the material of the white decoration. The results show that the raw material of the white barbotine decoration is white clay. Two types of applications can be distinguished: one method was to fire the barbotine together with the vessel at a temperature of about 950 °C. This method is characteristic for the majority of the barbotine decoration on Samian wares from Rheinzabern. In the second method the raw white clay was first fired at a high temperature (about 1100 °C) and then crushed. Water and perhaps potash-rich materials were added to this mixture to give it the proper consistency and plasticity to draw motifs on the surface of the coated vessels. This technology could be observed on some Samian wares from Rheinzabern and on all of the motto beakers from Trier. © 2013 Elsevier B.V. All rights reserved.
1. Introduction and archaeological background A special use of the rare white clay in the Roman time was the production of white barbotine decoration on quality pottery with black coat or red slip. This usage was applied on three main pottery groups. The first group can mainly be found in Central Gaul, in the surroundings of Lezoux. The vessels are generally beakers; the white barbotine decoration is applied on their black coat. The chronology of this group is only roughly known, they were probably made in the second half of the 2nd and in the first half of the 3rd centuries AD (Symonds, 1992). The second one is a group of the Samian ware from Rheinzabern. The form spectrum of the red slipped vessels with white barbotine decoration is broad: plates, beakers, jugs, etc. The precise chronology of this special group of Samian wares from Rheinzabern is also uncertain. Their production took place maybe at the end of the 2nd and in the first half of the 3rd centuries AD (Thomas, 2001). The third one is the group of the so called motto beakers from Trier. These were mostly drinking vessels, i.e. beakers, cups, jugs and flagons. This group was also black coated (similarly to the first group from Central Gaul); the white decoration was applied onto this coat. Of the three groups only the chronology of the motto beakers of Trier is known in ⁎ Corresponding author. Tel.: +36 3722500. E-mail addresses: [email protected] (E. Harsányi), [email protected] (I. Sajó), [email protected] (G. Szakmány), [email protected] (Z. Bendő).
detail, it was produced from 235/240 till 355 AD. (Harsányi, 2011; Künzl, 1997). Based on this chronological order it can be supposed, that the technique of the white decoration came first from Central Gaul to Rheinzabern and from there to Trier. 2. Aim The white barbotine decoration has been not analysed yet with natural scientific methods. A doctoral research on the black coated ware from Trier raised the question of the origin and the technology of the white barbotine decoration of these vessels and offered an opportunity to analyse and compare the white decoration of the motto beakers from Trier and of the Samian ware from Rheinzabern (Harsányi, 2011). The main questions were the following: 1) Of which components is the raw material of the white barbotine decoration composed? 2) How was it applied on the surface of the pottery? 3) At which temperature was it fired? Based on the comparison of the samples from two of the three pottery groups (motto beakers from Trier, Samian wares from Rheinzabern) we wanted to find out, how uniform the technology of the white barbotine decoration in the various workshops was.
0169-1317/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.clay.2013.06.028
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
Applied Clay Science xxx (2013) xxx–xxx
3. Methods
4.2. Microscopic petrography and SEM-EDS
Two samples from Trier (inv. no. KSZV 2004.I13/H13.103.12 and KSZV 2002.G12.046.98) and 4 samples from Rheinzabern (inv. no. E84/80/297, E82/53/196, E83/84/3 and E92/167) were investigated by petrographic and SEM-EDS analyses. In addition 12 XRD analyses, 3 samples from Trier (inv. no. KSZV 2004.I13/H13.103.12, KSZV 2002.G12.046.98 and BTM 51163) and 9 samples from Rheinzabern (inv. no. E84/80/297, E82/53/196, E82/84/3, E92/167, E84/50/716, E85/41/44, BTM 42295, BTM 55.28.43 and BTM without no.) were made. Macroscopically all the motto beakers from Trier with white barbotine seem to be produced with the same technique. The XRDanalyses of the samples with fast identical results proved this observation, therefore the small number of samples can be considered as representative (Table 1). SEM examinations and electron microprobe analyses were performed at the Department of Petrology and Geochemistry, Eötvös Loránd University using an AMRAY 1830 scanning electron microscope equipped with an EDAX PV 9800 EDS detector. Conditions of analysis: accelerating voltage 20 kV, beam current 1 nA, measurement time 100 s (livetime), beam diameter ~ 100 nm. Analyses were made on polished surfaces perpendicular to the wall of ceramics. X-ray powder diffraction scans were taken on a Philips PW 1050 diffractometer with CuKα radiation. Specimens were prepared from a few mg of thoroughly ground sample on a zero background sample holder. Scans were evaluated for quantitative phase composition using full profile fitting methods (Table 2).
The composition and appearance of the studied barbotines are very similar to each other, however, there are differences in texture and stucture. Generally there is a large amount of clasts in a variously vitrified clayey matrix in all the samples. The size of the clasts are variable, the dominant range of them is 1-50 μm, except for the all Trier and one Rheinzabern (E83/84/3) samples in which the clasts are 20–200 μm (Figs. 3B, C, D, 4B, C, D). Among the clasts quartz is predominant, there is less K-feldspar and very rarely relicts of 10 Å-phases can be found. Quartz and K-feldspar are initially resorbed in samples coming from Rheinzabern, while they are angular in samples from Trier. There is quite a large amount of rutile (1–5 μm angular clasts) accompanied by other accessories such as zircon and monazite. The amount of glass present varies, generally the highest amount is in the interface of the slip and the barbotine, however in some cases even the barbotine is completely glassy. Between the barbotine and the ceramic body a 10–20 μm thick slip can be observed, containing a considerable amount of submicron scale Fe-rich phase. Generally the slip and the barbotine are tightly connected due to a shallow deep fusion on the border of the barbotine and the slip. Except for the above mentioned features there are huge differences among the samples. Barbotine of the sample E84/80/297 is strongly and almost completely vitrified (Fig. 1B and C), while barbotine in sample E82/53/196 consists of two layers, the inner being less glassy than the outer part. The differences between the two layers manifest only in their composition, the transition zone is thin, but continuous. The strongly glassy outer part is richer in potassium than the inner layer probably producing completely glassy phases (K2O content are 5% and 13–17% respectively) (Fig. 1E). Sample E92/167 is the least glassy among the studied samples. Even the relicts of micas also can be recognized (Fig. 2C). In this sample the slip is ragged (Fig. 2B and D), probably because the barbotine was applied on not a completely consolidated surface. In sample E83/84/3 clasts are less well sorted and there are considerably coarser than in the other studied Rheinzabern samples and similar to that of the barbotine of the Trier samples (20–200 μm). In samples from Trier not only the size, but also their shape is different being more angular (Fig. 4C and D). In the barbotine of Trier pottery very fine (5–10 × 1 μm) needles can be seen, which are too thin to be analysed by SEM-EDS, but were identified as mullite by XRD analysis.
4. Results 4.1. Macroscopic petrography The pattern of the white decoration of the motto beakers from Trier is generally composed of a short motto and plant motifs underneath (Figs. 1A and D, 2A, 3A, 4A). The details of the compositional structure of the Samian wares with white decoration from Rheinzabern have not been published yet. It is known that a motto on the vessels is not as frequent as on the motto beakers from Trier, plant motifs rather dominate in different compositions. The barbotine on the motto beakers from Trier and on some Samian wares from Rheinzabern is snow-white, compact, but relatively soft; it can be easily detached from the surface. In the case of many motto beakers from Trier it even flakes off by itself. On another group of the Samian ware from Rheinzabern the barbotine is yellowish, its consistence is hard and it looks similar to the slip on the surface of the pottery. The barbotine can only be removed from these vessels together with the slip.
4.3. XRD Based on their X-ray powder diffraction patterns, all the investigated barbotine samples are characterized by high amount of quartz
Table 1 List of the analysed samples. Inv. no.
Pottery type
Colour of the coat
XRD
Petrographic and SEM-EDS analyses
XRD barbotine-type
Technological group
KSZV 2004.I13/H13.103.12 KSZV 2002.G12.046.98 BTM 51163 E84/80/297 E82/53/196 E83/84/3 E92/167 E84/50/716 E85/41/44 BTM 42295 BTM 55.28.43
motto beaker from Trier motto beaker from Trier motto beaker from Trier Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern
black black black red red red red red red red red
yes yes yes yes yes yes yes yes yes yes yes
yes yes no yes yes yes yes no no no no
group b group b
BTM without no.
Samian ware from Rheinzabern
red
yes
no
type 3 type 3 type 3 type 1 ? type 2 type 2 type 1 type 2 type 2 type 2? (poor XRD) type 2
group b group b group b group a not determined not determined not determined not determined not determined
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
3
Table 2 Quantitative phase composition of the three different barbotine-types. Barbotine type
Type 1 Type 2 Type 3
Barbotine sample
E84/50/716, Rheinzabern E83/84/3, Rheinzabern KSZV 2002.G12.046.98, Trier
Phase composition (w/w percentages) Glass
Quartz
Illite
Anatase
Rutil
Mullite
Corundum
40 40 40
30 50 50
20 – –
1 – 0.5
2 1 2
– 5 5
– 4 –
and the lack of iron containing phases, which explains the white colour of the decoration. Furthermore, the barbotine samples from Trier exhibit a quite uniform mineralogy. On the other hand, samples from Rheinzabern show a greater variety in mineralogical composition, and can be divided roughly into two groups, both different from the Trier type. These three technological–mineralogical types can be characterized as follows:
As regards the correlation between vitrification degree and XRD results, the amount of the glassy phases of the analysed barbotine is 25–40% and depends on the K-content. We did not detect any high temperature minerals by XRD (e.g. mullite, corundum) but there are notable 10 Å phases in the barbotines with high K-content (K2O content N 10 %). On the contrary there are mullite and/or corundum in the barbotines with less than 10% K2O-content. 5. Discussion and conclusion
type 1: a group of barbotine samples from Rheinzabern (E84/80/297 and E84/50/716) contains a fair amount of illite, accompanied with rutile and anatase. No high temperature phases (mullite, corundum) can be detected. Illite residues indicate a firing temperature of below 950 °C (Fig. 5 scan A). type 2: another group of barbotine samples from Rheinzabern (E85/41/ 44, E83/84/3, E92/167 and BTM without inv. no.) contains rutile but no anatase is detected. Clay residues are lacking and mullite and corundum are formed in the firing process. The presence of corundum refers to a firing temperature around or above 1100 °C (Fig. 5 scan B). Also the lack of anatase shows a high temperature. type 3: in the barbotine samples from Trier (KSZV 2004.I13/H13.103.12, KSZV 2002.G12.046.98 and BTM 51163) the dominant quartz is accompanied with rutile and occasionally some anatase. Mullite is always present and no residues of clay minerals can be detected (Fig. 5 scan C). The presence of mullite and the lack of clay residues refer to a firing temperature above 1000 °C.
The analyses have shown that the material and the technology of the white barbotine decoration were not uniform. The following two technological groups (a) and (b) can be distinguished: (a) By this method white clay, poor in iron, was probably mixed with quartz sand or quartzite grit, which contains rutile, zircon and little monazite, or originally it contained fine grained quartz. This mixture had to get the proper consistence and plasticity to be applied in one or more layers onto the surface of the coated but not yet fired vessels. We suppose that in the case of more layers the higher potassium content in the outer part might be due to the use of more levigated clay. With this method there was only one firing: the barbotine was fired together with the vessel at a temperature of about 950 °C. While only one of the fully characterized samples was assigned to group a, macroscopic observations as indicated above (e.g. hardness of decoration, barbotine can only be removed together with slip) suggest
Fig. 1. Strongly vitrified barbotine, with few initially resorbed coarse grained quartz remnants, and some fine grained zircon and rutile. There is a thin (10–15 μm) iron rich slip layer of uniform thickness between the ceramic body and the barbotine. A, B, C: glassy barbotine forming one layer (sample E84/80/297) D, E, F: glassy barbotine forming two layers: the rim is more vitrified due to its higher K-content (sample E82/53/196).
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
4
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
Fig. 2. Non-vitrified barbotine with quartz, 10 Å phyllosilicate and feldspar remnants. There is a soft layer between the barbotine and the ceramic body which is not continuous: the material of the rim is sometimes torn and the broken parts are thrusted. The presence of this soft layer between the ceramic body and the barbotine suggests that the ceramic was not fired before putting the barbotine (sample E92/167) on.
that this method is characteristic for the main part of barbotine decoration on Samian wares from Rheinzabern. (b) In the case of the second method the raw material was white clay, poor in iron. On the basis of the results of the analyses we think that the clays was first fired at a high temperature (about 1100 °C), then crushed and mixed additionally with quartz-sand or quartzite grit, which contains rutile, zircon and monazite.
Adding water and supposedly potash to this mixture it got the proper consistence and plasticity to draw motifs on the surface of the coated vessels. We cannot answer yet whether the vessel was already fired at a temperature of about 950 °C as our and other XRD-analyses showed (Bocquet, 1999; Harsányi, 2011), and the barbotine was fired onto the surface on a lower temperature, or whether the white decoration was applied onto the
Fig. 3. Poorly vitrified barbotine on terrra sigillata. The vitrification in the barbotine is more intensive near the thin, uniform iron-rich slip layer. The barbotine contains weakly resorbed quartz, 10 Å phyllosilicate, and quite a lot of fine grained zircon, rutile and few monazite (sample E83/84/3).
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
5
Fig. 4. Poorly vitrified barbotine on Trier ceramic. The vitrification in the barbotine is more intensive near the thin, uniform iron-rich slip layer. The barbotine contains weakly resorbed quartz, 10 Å phyllosilicate, and quite a lot of fine grained zircon, rutile and few monazite (sample KSZV 2002.G12.046.98).
surface of unfired vessels and both were fired together. An argument for the double firing is that the white barbotine — mainly on the motto beakers from Trier — often flakes off from the surface without damaging the coat. Thus it seems that the two surfaces were not “melted” during the firing. An argument for the simple firing method is that the double firing has been very expensive and there is no direct archaeological evidence for it in the Roman time. But it should be noticed that, based on the archaeological contexts, the motto beakers from Trier were expensive vessels (Harsányi, 2011). So it seems that there was a clientele for who it was not a problem to pay a higher price for a decorative quality pottery.
This technology could be observed on some Samian wares from Rheinzabern and on all of the motto beakers from Trier. This method was presumably the newer and/or better one, developed in Rheinzabern, so it seems that only this one was borrowed and used in Trier for the decoration on the black coated ware. Due to the small number of the samples these results are indicative. The precise provenance of the white raw clay used for the barbotine decoration has not been discovered yet. In the case of method (b) it can be assumed that sherds of broken white vessels, fired originally to high temperature, were crushed and used to produce the mixture for the barbotine decoration. The nearest find place of white raw clay to Trier is about 20 km north, in the surrounding of Speicher. This raw clay
Fig. 5. Three different XRD patterns of barbotine (see details in the text). Abbreviations: a — anatase; i — illite; q — quartz; m — mullite; k — corundum; r — rutile.
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
6
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
was known and used also in the Roman period (Gilles, 1983), thus the raw material to the barbotine decoration could be extracted from this place or white broken sherds from here could be used for the raw material of the white barbotine. Other possible find spots of raw clay or for pottery industry are for example Cologne, Bonn or Urmitz, where white clay could be found and white vessels, fired to high temperature were produced (Friedrich, 2011; Höpken, 2005). In the areas near Lezoux white clay can be found in the valley of the river Allier. It was also used in the Roman times, mainly for production of terracotta figurines (Lahanier and Rouvier-Jeanlin, 1977). Unfortunately we have not found any data yet about the composition of the raw clays of all these sites in the Roman and/or in the modern times. To answer the question of the origin of the material of the white barbotine, samples of vessels and raw clay from these places must be analysed. In the surroundings of Rheinzabern no find spot of white raw clay is known.
Jean-Frank Wagner (Departement of Geology, University of Trier) for his help to find the spots for white raw clay in the surroundings of Speicher. Dr. Constanze Höpken (Departement of Archaeology, University of Cologne) gave us useful information about the details of the production of Roman potteries. We thank Sándor Józsa (ELTE Department of Petrology and Geochemistry, Budapest) for his excellent work in preparing the samples.
6. Future plan
References
Samples of vessels with white decoration from Central Gaul have to be analysed to completely answer the question of the origin and the technology of the white decoration. The search for such samples is in progress. Besides, we plan to make test firings of white raw clay from Speicher and to analyse these fired samples and Roman white sherds from this place. We expect that it will become clear whether Speicher can be the potential original place of the raw material for the white barbotine or whether we should continue to investigate the origin of the raw clay of the decoration of these potteries. Acknowledgements We would like to thank Dr. Annamária R. Facsády and Dr. Paula Zsidi (Aquincum Museum, Budapest) together with Dr. Rüdiger Schulz (Generaldirekation Kulturelles Erbe Rheinland-Pfalz, Speyer) and Dr. Bartus Dávid (ELTE, Institute of Archaeology, Budapest) to make possible the analyses of motto beakers from Trier and Samian wares with white decoration from Rheinzabern. We thank Prof. Dr.
Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at http://dx.doi.org/10.1016/j.clay.2013.06.028. These data include Google maps of the most important areas described in this article.
Bocquet, A., 1999. La production et la distribution des céramiques fines engobées et métallescentes dans le nord de la Gaule: approche minéralogique et géochimique. In: Brulet, R., Symonds, R.P., Vilvorder, F. (Eds.), Céramiques engobées et métallescentes gallo-romaines. Actes du colloque organisé à Louvain-la-Neuve le 18 mars 1995, Oxford, pp. 129–288. Gilles, K.-J., 1983. Speicherer Keramik. In: Cüppers, H. (Ed.), Die Römer an Mosel und Saar. Zeugnisse der Römerzeit in Lothringen, in Luxemburg, im Raum Trier und im Saarland, Bonn, pp. 214–218. Friedrich, S., 2011. Die römischen Töpfereien von Weissenthurm. Limes 5 (1), 19–23. Harsányi, E., 2011. Die Trierer schwarz engobierte Ware und ihre Imitationen in Noricum und Pannonien (Unpubl. doctoral thesis, Köln 684 pp.). Höpken, C., 2005. Die römische Keramikproduktion in Köln. Kölner Forschungen 8, Köln. (660 pp.). Künzl, S., 1997. Die Trierer Spruchbecherkeramik. Dekorierte Schwarzfirniskeramik des 3. und 4. Jahrhunderts n. Chr. Trierer Zeitschrift (Beiheft 21) (Trier, 379 pp.). Lahanier, Ch., Rouvier-Jeanlin, M., 1977. Analyses de 120 figurines gallo-romaines en terre cuite blanche. In: Hours, M. (Ed.), L‘analyse par microfluorescence X appliquée à l’archéologie. Actes du Colloque organisé au Laboratoire des Musées de France en juin 1977. PACT (Journal of the European Study Group on Physical, Chemical and Mathematical Techniques Applied to Archaeology), 1, pp. 110–130. Symonds, R.P., 1992. Rhenish Wares. Fine Dark Coloured Pottery from Gaul and Germany. Monograph No. 23. Oxford University Committee for Archaeology, Oxford (180 pp.). Thomas, M., 2001. Terra Sigillata mit Weissbarbotine-Verzierung aus Rheinzabern. Rei Cretariae Romanae Fautores Acta 37, 243–245.
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
1
Contents lists available at SciVerse ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Analyses of the white barbotine decoration of two Roman pottery groups Eszter Harsányi a, István Sajó b, György Szakmány c,⁎, Zsolt Bendő c a b c
Archäologisches Landesmuseum Baden-Württemberg, Außenstelle Rastatt, Lützowerstraße 10, 76437 Rastatt, Germany Institute of Materials and Environmental Chemistry, Research Center for Natural Sciences, Hungarian Academy of Sciences, H-1025 Budapest, Pusztaszeri út 59-67, Hungary Department of Petrology and Geochemistry, Eötvös Loránd University, H-1117 Budapest, Pázmány P. sétány 1/c, Hungary
a r t i c l e
i n f o
Article history: Received 26 September 2012 Received in revised form 20 June 2013 Accepted 26 June 2013 Available online xxxx Keywords: White barbotine Roman pottery Black coated ware Samian Ware Trier Rheinzabern
a b s t r a c t White barbotine is a rare decoration form in Roman pottery production. It was applied on three main pottery groups from the 2nd to the 4th century AD: on the Samian ware from Rheinzabern, on the black coated ware from central Gaul and on the motto beakers from Trier. This kind of decoration has not yet been analysed with scientific methods. As a part of a research on the black coated ware in Pannonia some motto beakers from Trier and some samples of Samian ware from Rheinzabern were analysed with XRD, petrographic and EDS-EDX methods to get information about the technique and the material of the white decoration. The results show that the raw material of the white barbotine decoration is white clay. Two types of applications can be distinguished: one method was to fire the barbotine together with the vessel at a temperature of about 950 °C. This method is characteristic for the majority of the barbotine decoration on Samian wares from Rheinzabern. In the second method the raw white clay was first fired at a high temperature (about 1100 °C) and then crushed. Water and perhaps potash-rich materials were added to this mixture to give it the proper consistency and plasticity to draw motifs on the surface of the coated vessels. This technology could be observed on some Samian wares from Rheinzabern and on all of the motto beakers from Trier. © 2013 Elsevier B.V. All rights reserved.
1. Introduction and archaeological background A special use of the rare white clay in the Roman time was the production of white barbotine decoration on quality pottery with black coat or red slip. This usage was applied on three main pottery groups. The first group can mainly be found in Central Gaul, in the surroundings of Lezoux. The vessels are generally beakers; the white barbotine decoration is applied on their black coat. The chronology of this group is only roughly known, they were probably made in the second half of the 2nd and in the first half of the 3rd centuries AD (Symonds, 1992). The second one is a group of the Samian ware from Rheinzabern. The form spectrum of the red slipped vessels with white barbotine decoration is broad: plates, beakers, jugs, etc. The precise chronology of this special group of Samian wares from Rheinzabern is also uncertain. Their production took place maybe at the end of the 2nd and in the first half of the 3rd centuries AD (Thomas, 2001). The third one is the group of the so called motto beakers from Trier. These were mostly drinking vessels, i.e. beakers, cups, jugs and flagons. This group was also black coated (similarly to the first group from Central Gaul); the white decoration was applied onto this coat. Of the three groups only the chronology of the motto beakers of Trier is known in ⁎ Corresponding author. Tel.: +36 3722500. E-mail addresses: [email protected] (E. Harsányi), [email protected] (I. Sajó), [email protected] (G. Szakmány), [email protected] (Z. Bendő).
detail, it was produced from 235/240 till 355 AD. (Harsányi, 2011; Künzl, 1997). Based on this chronological order it can be supposed, that the technique of the white decoration came first from Central Gaul to Rheinzabern and from there to Trier. 2. Aim The white barbotine decoration has been not analysed yet with natural scientific methods. A doctoral research on the black coated ware from Trier raised the question of the origin and the technology of the white barbotine decoration of these vessels and offered an opportunity to analyse and compare the white decoration of the motto beakers from Trier and of the Samian ware from Rheinzabern (Harsányi, 2011). The main questions were the following: 1) Of which components is the raw material of the white barbotine decoration composed? 2) How was it applied on the surface of the pottery? 3) At which temperature was it fired? Based on the comparison of the samples from two of the three pottery groups (motto beakers from Trier, Samian wares from Rheinzabern) we wanted to find out, how uniform the technology of the white barbotine decoration in the various workshops was.
0169-1317/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.clay.2013.06.028
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
Applied Clay Science xxx (2013) xxx–xxx
3. Methods
4.2. Microscopic petrography and SEM-EDS
Two samples from Trier (inv. no. KSZV 2004.I13/H13.103.12 and KSZV 2002.G12.046.98) and 4 samples from Rheinzabern (inv. no. E84/80/297, E82/53/196, E83/84/3 and E92/167) were investigated by petrographic and SEM-EDS analyses. In addition 12 XRD analyses, 3 samples from Trier (inv. no. KSZV 2004.I13/H13.103.12, KSZV 2002.G12.046.98 and BTM 51163) and 9 samples from Rheinzabern (inv. no. E84/80/297, E82/53/196, E82/84/3, E92/167, E84/50/716, E85/41/44, BTM 42295, BTM 55.28.43 and BTM without no.) were made. Macroscopically all the motto beakers from Trier with white barbotine seem to be produced with the same technique. The XRDanalyses of the samples with fast identical results proved this observation, therefore the small number of samples can be considered as representative (Table 1). SEM examinations and electron microprobe analyses were performed at the Department of Petrology and Geochemistry, Eötvös Loránd University using an AMRAY 1830 scanning electron microscope equipped with an EDAX PV 9800 EDS detector. Conditions of analysis: accelerating voltage 20 kV, beam current 1 nA, measurement time 100 s (livetime), beam diameter ~ 100 nm. Analyses were made on polished surfaces perpendicular to the wall of ceramics. X-ray powder diffraction scans were taken on a Philips PW 1050 diffractometer with CuKα radiation. Specimens were prepared from a few mg of thoroughly ground sample on a zero background sample holder. Scans were evaluated for quantitative phase composition using full profile fitting methods (Table 2).
The composition and appearance of the studied barbotines are very similar to each other, however, there are differences in texture and stucture. Generally there is a large amount of clasts in a variously vitrified clayey matrix in all the samples. The size of the clasts are variable, the dominant range of them is 1-50 μm, except for the all Trier and one Rheinzabern (E83/84/3) samples in which the clasts are 20–200 μm (Figs. 3B, C, D, 4B, C, D). Among the clasts quartz is predominant, there is less K-feldspar and very rarely relicts of 10 Å-phases can be found. Quartz and K-feldspar are initially resorbed in samples coming from Rheinzabern, while they are angular in samples from Trier. There is quite a large amount of rutile (1–5 μm angular clasts) accompanied by other accessories such as zircon and monazite. The amount of glass present varies, generally the highest amount is in the interface of the slip and the barbotine, however in some cases even the barbotine is completely glassy. Between the barbotine and the ceramic body a 10–20 μm thick slip can be observed, containing a considerable amount of submicron scale Fe-rich phase. Generally the slip and the barbotine are tightly connected due to a shallow deep fusion on the border of the barbotine and the slip. Except for the above mentioned features there are huge differences among the samples. Barbotine of the sample E84/80/297 is strongly and almost completely vitrified (Fig. 1B and C), while barbotine in sample E82/53/196 consists of two layers, the inner being less glassy than the outer part. The differences between the two layers manifest only in their composition, the transition zone is thin, but continuous. The strongly glassy outer part is richer in potassium than the inner layer probably producing completely glassy phases (K2O content are 5% and 13–17% respectively) (Fig. 1E). Sample E92/167 is the least glassy among the studied samples. Even the relicts of micas also can be recognized (Fig. 2C). In this sample the slip is ragged (Fig. 2B and D), probably because the barbotine was applied on not a completely consolidated surface. In sample E83/84/3 clasts are less well sorted and there are considerably coarser than in the other studied Rheinzabern samples and similar to that of the barbotine of the Trier samples (20–200 μm). In samples from Trier not only the size, but also their shape is different being more angular (Fig. 4C and D). In the barbotine of Trier pottery very fine (5–10 × 1 μm) needles can be seen, which are too thin to be analysed by SEM-EDS, but were identified as mullite by XRD analysis.
4. Results 4.1. Macroscopic petrography The pattern of the white decoration of the motto beakers from Trier is generally composed of a short motto and plant motifs underneath (Figs. 1A and D, 2A, 3A, 4A). The details of the compositional structure of the Samian wares with white decoration from Rheinzabern have not been published yet. It is known that a motto on the vessels is not as frequent as on the motto beakers from Trier, plant motifs rather dominate in different compositions. The barbotine on the motto beakers from Trier and on some Samian wares from Rheinzabern is snow-white, compact, but relatively soft; it can be easily detached from the surface. In the case of many motto beakers from Trier it even flakes off by itself. On another group of the Samian ware from Rheinzabern the barbotine is yellowish, its consistence is hard and it looks similar to the slip on the surface of the pottery. The barbotine can only be removed from these vessels together with the slip.
4.3. XRD Based on their X-ray powder diffraction patterns, all the investigated barbotine samples are characterized by high amount of quartz
Table 1 List of the analysed samples. Inv. no.
Pottery type
Colour of the coat
XRD
Petrographic and SEM-EDS analyses
XRD barbotine-type
Technological group
KSZV 2004.I13/H13.103.12 KSZV 2002.G12.046.98 BTM 51163 E84/80/297 E82/53/196 E83/84/3 E92/167 E84/50/716 E85/41/44 BTM 42295 BTM 55.28.43
motto beaker from Trier motto beaker from Trier motto beaker from Trier Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern Samian ware from Rheinzabern
black black black red red red red red red red red
yes yes yes yes yes yes yes yes yes yes yes
yes yes no yes yes yes yes no no no no
group b group b
BTM without no.
Samian ware from Rheinzabern
red
yes
no
type 3 type 3 type 3 type 1 ? type 2 type 2 type 1 type 2 type 2 type 2? (poor XRD) type 2
group b group b group b group a not determined not determined not determined not determined not determined
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
3
Table 2 Quantitative phase composition of the three different barbotine-types. Barbotine type
Type 1 Type 2 Type 3
Barbotine sample
E84/50/716, Rheinzabern E83/84/3, Rheinzabern KSZV 2002.G12.046.98, Trier
Phase composition (w/w percentages) Glass
Quartz
Illite
Anatase
Rutil
Mullite
Corundum
40 40 40
30 50 50
20 – –
1 – 0.5
2 1 2
– 5 5
– 4 –
and the lack of iron containing phases, which explains the white colour of the decoration. Furthermore, the barbotine samples from Trier exhibit a quite uniform mineralogy. On the other hand, samples from Rheinzabern show a greater variety in mineralogical composition, and can be divided roughly into two groups, both different from the Trier type. These three technological–mineralogical types can be characterized as follows:
As regards the correlation between vitrification degree and XRD results, the amount of the glassy phases of the analysed barbotine is 25–40% and depends on the K-content. We did not detect any high temperature minerals by XRD (e.g. mullite, corundum) but there are notable 10 Å phases in the barbotines with high K-content (K2O content N 10 %). On the contrary there are mullite and/or corundum in the barbotines with less than 10% K2O-content. 5. Discussion and conclusion
type 1: a group of barbotine samples from Rheinzabern (E84/80/297 and E84/50/716) contains a fair amount of illite, accompanied with rutile and anatase. No high temperature phases (mullite, corundum) can be detected. Illite residues indicate a firing temperature of below 950 °C (Fig. 5 scan A). type 2: another group of barbotine samples from Rheinzabern (E85/41/ 44, E83/84/3, E92/167 and BTM without inv. no.) contains rutile but no anatase is detected. Clay residues are lacking and mullite and corundum are formed in the firing process. The presence of corundum refers to a firing temperature around or above 1100 °C (Fig. 5 scan B). Also the lack of anatase shows a high temperature. type 3: in the barbotine samples from Trier (KSZV 2004.I13/H13.103.12, KSZV 2002.G12.046.98 and BTM 51163) the dominant quartz is accompanied with rutile and occasionally some anatase. Mullite is always present and no residues of clay minerals can be detected (Fig. 5 scan C). The presence of mullite and the lack of clay residues refer to a firing temperature above 1000 °C.
The analyses have shown that the material and the technology of the white barbotine decoration were not uniform. The following two technological groups (a) and (b) can be distinguished: (a) By this method white clay, poor in iron, was probably mixed with quartz sand or quartzite grit, which contains rutile, zircon and little monazite, or originally it contained fine grained quartz. This mixture had to get the proper consistence and plasticity to be applied in one or more layers onto the surface of the coated but not yet fired vessels. We suppose that in the case of more layers the higher potassium content in the outer part might be due to the use of more levigated clay. With this method there was only one firing: the barbotine was fired together with the vessel at a temperature of about 950 °C. While only one of the fully characterized samples was assigned to group a, macroscopic observations as indicated above (e.g. hardness of decoration, barbotine can only be removed together with slip) suggest
Fig. 1. Strongly vitrified barbotine, with few initially resorbed coarse grained quartz remnants, and some fine grained zircon and rutile. There is a thin (10–15 μm) iron rich slip layer of uniform thickness between the ceramic body and the barbotine. A, B, C: glassy barbotine forming one layer (sample E84/80/297) D, E, F: glassy barbotine forming two layers: the rim is more vitrified due to its higher K-content (sample E82/53/196).
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
4
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
Fig. 2. Non-vitrified barbotine with quartz, 10 Å phyllosilicate and feldspar remnants. There is a soft layer between the barbotine and the ceramic body which is not continuous: the material of the rim is sometimes torn and the broken parts are thrusted. The presence of this soft layer between the ceramic body and the barbotine suggests that the ceramic was not fired before putting the barbotine (sample E92/167) on.
that this method is characteristic for the main part of barbotine decoration on Samian wares from Rheinzabern. (b) In the case of the second method the raw material was white clay, poor in iron. On the basis of the results of the analyses we think that the clays was first fired at a high temperature (about 1100 °C), then crushed and mixed additionally with quartz-sand or quartzite grit, which contains rutile, zircon and monazite.
Adding water and supposedly potash to this mixture it got the proper consistence and plasticity to draw motifs on the surface of the coated vessels. We cannot answer yet whether the vessel was already fired at a temperature of about 950 °C as our and other XRD-analyses showed (Bocquet, 1999; Harsányi, 2011), and the barbotine was fired onto the surface on a lower temperature, or whether the white decoration was applied onto the
Fig. 3. Poorly vitrified barbotine on terrra sigillata. The vitrification in the barbotine is more intensive near the thin, uniform iron-rich slip layer. The barbotine contains weakly resorbed quartz, 10 Å phyllosilicate, and quite a lot of fine grained zircon, rutile and few monazite (sample E83/84/3).
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
5
Fig. 4. Poorly vitrified barbotine on Trier ceramic. The vitrification in the barbotine is more intensive near the thin, uniform iron-rich slip layer. The barbotine contains weakly resorbed quartz, 10 Å phyllosilicate, and quite a lot of fine grained zircon, rutile and few monazite (sample KSZV 2002.G12.046.98).
surface of unfired vessels and both were fired together. An argument for the double firing is that the white barbotine — mainly on the motto beakers from Trier — often flakes off from the surface without damaging the coat. Thus it seems that the two surfaces were not “melted” during the firing. An argument for the simple firing method is that the double firing has been very expensive and there is no direct archaeological evidence for it in the Roman time. But it should be noticed that, based on the archaeological contexts, the motto beakers from Trier were expensive vessels (Harsányi, 2011). So it seems that there was a clientele for who it was not a problem to pay a higher price for a decorative quality pottery.
This technology could be observed on some Samian wares from Rheinzabern and on all of the motto beakers from Trier. This method was presumably the newer and/or better one, developed in Rheinzabern, so it seems that only this one was borrowed and used in Trier for the decoration on the black coated ware. Due to the small number of the samples these results are indicative. The precise provenance of the white raw clay used for the barbotine decoration has not been discovered yet. In the case of method (b) it can be assumed that sherds of broken white vessels, fired originally to high temperature, were crushed and used to produce the mixture for the barbotine decoration. The nearest find place of white raw clay to Trier is about 20 km north, in the surrounding of Speicher. This raw clay
Fig. 5. Three different XRD patterns of barbotine (see details in the text). Abbreviations: a — anatase; i — illite; q — quartz; m — mullite; k — corundum; r — rutile.
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028
6
E. Harsányi et al. / Applied Clay Science xxx (2013) xxx–xxx
was known and used also in the Roman period (Gilles, 1983), thus the raw material to the barbotine decoration could be extracted from this place or white broken sherds from here could be used for the raw material of the white barbotine. Other possible find spots of raw clay or for pottery industry are for example Cologne, Bonn or Urmitz, where white clay could be found and white vessels, fired to high temperature were produced (Friedrich, 2011; Höpken, 2005). In the areas near Lezoux white clay can be found in the valley of the river Allier. It was also used in the Roman times, mainly for production of terracotta figurines (Lahanier and Rouvier-Jeanlin, 1977). Unfortunately we have not found any data yet about the composition of the raw clays of all these sites in the Roman and/or in the modern times. To answer the question of the origin of the material of the white barbotine, samples of vessels and raw clay from these places must be analysed. In the surroundings of Rheinzabern no find spot of white raw clay is known.
Jean-Frank Wagner (Departement of Geology, University of Trier) for his help to find the spots for white raw clay in the surroundings of Speicher. Dr. Constanze Höpken (Departement of Archaeology, University of Cologne) gave us useful information about the details of the production of Roman potteries. We thank Sándor Józsa (ELTE Department of Petrology and Geochemistry, Budapest) for his excellent work in preparing the samples.
6. Future plan
References
Samples of vessels with white decoration from Central Gaul have to be analysed to completely answer the question of the origin and the technology of the white decoration. The search for such samples is in progress. Besides, we plan to make test firings of white raw clay from Speicher and to analyse these fired samples and Roman white sherds from this place. We expect that it will become clear whether Speicher can be the potential original place of the raw material for the white barbotine or whether we should continue to investigate the origin of the raw clay of the decoration of these potteries. Acknowledgements We would like to thank Dr. Annamária R. Facsády and Dr. Paula Zsidi (Aquincum Museum, Budapest) together with Dr. Rüdiger Schulz (Generaldirekation Kulturelles Erbe Rheinland-Pfalz, Speyer) and Dr. Bartus Dávid (ELTE, Institute of Archaeology, Budapest) to make possible the analyses of motto beakers from Trier and Samian wares with white decoration from Rheinzabern. We thank Prof. Dr.
Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version, at http://dx.doi.org/10.1016/j.clay.2013.06.028. These data include Google maps of the most important areas described in this article.
Bocquet, A., 1999. La production et la distribution des céramiques fines engobées et métallescentes dans le nord de la Gaule: approche minéralogique et géochimique. In: Brulet, R., Symonds, R.P., Vilvorder, F. (Eds.), Céramiques engobées et métallescentes gallo-romaines. Actes du colloque organisé à Louvain-la-Neuve le 18 mars 1995, Oxford, pp. 129–288. Gilles, K.-J., 1983. Speicherer Keramik. In: Cüppers, H. (Ed.), Die Römer an Mosel und Saar. Zeugnisse der Römerzeit in Lothringen, in Luxemburg, im Raum Trier und im Saarland, Bonn, pp. 214–218. Friedrich, S., 2011. Die römischen Töpfereien von Weissenthurm. Limes 5 (1), 19–23. Harsányi, E., 2011. Die Trierer schwarz engobierte Ware und ihre Imitationen in Noricum und Pannonien (Unpubl. doctoral thesis, Köln 684 pp.). Höpken, C., 2005. Die römische Keramikproduktion in Köln. Kölner Forschungen 8, Köln. (660 pp.). Künzl, S., 1997. Die Trierer Spruchbecherkeramik. Dekorierte Schwarzfirniskeramik des 3. und 4. Jahrhunderts n. Chr. Trierer Zeitschrift (Beiheft 21) (Trier, 379 pp.). Lahanier, Ch., Rouvier-Jeanlin, M., 1977. Analyses de 120 figurines gallo-romaines en terre cuite blanche. In: Hours, M. (Ed.), L‘analyse par microfluorescence X appliquée à l’archéologie. Actes du Colloque organisé au Laboratoire des Musées de France en juin 1977. PACT (Journal of the European Study Group on Physical, Chemical and Mathematical Techniques Applied to Archaeology), 1, pp. 110–130. Symonds, R.P., 1992. Rhenish Wares. Fine Dark Coloured Pottery from Gaul and Germany. Monograph No. 23. Oxford University Committee for Archaeology, Oxford (180 pp.). Thomas, M., 2001. Terra Sigillata mit Weissbarbotine-Verzierung aus Rheinzabern. Rei Cretariae Romanae Fautores Acta 37, 243–245.
Please cite this article as: Harsányi, E., et al., Analyses of the white barbotine decoration of two Roman pottery groups, Applied Clay Science (2013), http://dx.doi.org/10.1016/j.clay.2013.06.028