Next Article in Journal
Uranium Occurrence State and Its Implication for Sandstone-Type Uranium Mineralization within the Hanbazhai Area of the Longchuanjiang Basin, China
Next Article in Special Issue
A Multianalytical Approach to Identifying the White Marbles Used in Roman Imperial Sculptures from Tarraco (Hispania)
Previous Article in Journal
Experimental Study on Pore Structure and Soil-Water Characteristic Curve of Ionic Rare Earth Ore under Seepage
Previous Article in Special Issue
The Carbon and Oxygen Isotope Composition of the Marble Inscriptions of Aléria, Corsica
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Trading Marble for Steel: Early Roman Import of Carrara Marble into the Alps—The Example of the Magdalensberg Trading Post in Noricum

Austrian Archaeological Institute, Austrian Academy of Sciences, Franz Klein-Gasse 1, 1190 Vienna, Austria
Minerals 2023, 13(8), 1036; https://doi.org/10.3390/min13081036
Submission received: 29 June 2023 / Revised: 21 July 2023 / Accepted: 26 July 2023 / Published: 2 August 2023

Abstract

:
The annexation of the Noricum Kingdom by the Roman Empire in 16 BC brought an increase in the trading relations between the empire and its northern neighbours. A first hub for these relations was the emporium on the Magdalensberg in Noricum Mediterraneum (today southern Carinthia/Austria). During the last decades, archaeological investigations of this settlement in a remote mountainous area revealed, inter alia, different kinds of marble decoration and architecture. Provenance analyses using a combination of different methods, including isotope analysis, trace element analysis and the analysis of inclusion fluids, show that the marbles used on the Magdalensberg are of different origins. Widely used were medium- to coarse-grained Alpine marbles from Roman quarries of the region of Gummern. Prominently used for plates, tiles, profiles, etc. were several types of fine-grained marbles of different origins. One group definitely originated from the quarries of Carrara (Carrara white and Carrara Bardiglio), testifying to the trading relations with northern Italy after the integration of Noricum into the Roman Empire. A database for the Carrara Bardiglio marble is presented and discussed. For the use of these data by further investigators, the numerical data are given as online material.

1. Introduction

A large number of deposits of raw materials triggered virulent La Tène Celtic mining activity in the region of today’s southern Carinthia in the Eastern Alps of Austria. The special chemical and mineralogical characteristics of the manganese-rich iron ores of the Erzberg of Hüttenberg required an elaborate method of processing. It was exactly for this reason that ingenious Celtic metallurgists and iron forgers developed a sophisticated and complex technology to produce a usable product out of this complex ore. They managed to produce steel-like iron of unprecedented quality, the famous “Ferrum Noricum” which was highly sought-after in Rome; already, Pliny in his “Naturalis Historia” (XXXIV, 145) had praised its outstanding quality [1,2,3].
In the vicinity of a Celtic mountain sanctuary, the emporium on the Magdalensberg developed in the 1st century BC, and soon, trading relations with the northward expanding Roman Empire (e.g., the foundation of Aquileia in the early 2nd century BC) were established. The original Celtic or Roman name of the settlement is not handed down. In this early period, the Roman demand was not due to political claim, but rather stalwart economic interest based on the above-mentioned production of raw materials. Finally, political developments led to the annexation of the Noricum Kingdom in 16 BC. In this time of increasing economic relations, the trading centre of the Magdalensberg, at an altitude of more than 900 m, prospered.
Based on the above-mentioned production of excellent iron, but also of other raw materials like gold and copper, this settlement became a regional trading centre. After many decades of close relations to the empire with relative independence, the region became the Roman province of Noricum under Claudius. This development called for a more suitable administrative centre in a more convenient logistic position. The position of the Magdalensberg in a remote mountainous area was not very favourable; thus, it lost its strategic sense in peaceful times, and before the middle of the 1st century AD, the former trading centre of the Magdalensberg was abruptly abandoned. The new centre was founded in the Glan valley some 10 km away in a straight line from the Magdalensberg. The “Municipium Claudium Virunum” was situated at a crossroads from the river Danube to Aquileia [4,5]. It is assumed by different scholars that bigger architectural parts and marble decorations were then transported from the Magdalensberg settlement to Virunum [6]. During the raids by the Huns in the 5th century, the unfortified town of Virunum was completely destroyed. In this paper, the marble remnants of the Magdalensberg emporium testifying to the trading relations with the Roman Empire are investigated. Not much is left of the original marble inventory except for small, heavily fragmented findings of marble artefacts found on the Magdalensberg after a few decades of blooming. The results of ongoing investigations on the marble inventory of Virunum will be published later.
Substantial archaeological work on the Magdalensberg, and to some extent also on the marbles used, has been performed so far; however, published analytical data are very rare. In a series of papers, Steiner presented the descriptions and the corresponding isotope data of a large number of artefacts [7,8,9,10].

2. The Different Types of Regionally Occurring Marbles

As shown in Figure 1, a series of marble quarries can be found in the area of southern Noricum. A description and chemical and petrographic data on several of the quarries in Noricum, which are relevant for the marbles used on the Magdalensberg, were published in 1999 [11]. Stable isotope data and also bulk chemical analyses (total digestion using HF) of the marbles from these quarries were presented in the above-mentioned paper. However, a direct comparison of the chemical analyses published there with the databank used in this work may be problematic because of slight methodological differences [12,13]. There, a detailed description is given, and therefore, in the following, only a condensed overview is presented.
Geologically, two different units in southern Noricum host several marble-bearing formations with traces of quarrying activities dating back already to Roman times. The Koralpe–Wölz System and the tectonically higher Drauzug–Gurktal System are two subunits of the Upper Austroalpine unit of the Eastern Alps. The marbles of these two formations can easily be distinguished with the naked eye because of the very different grain size due to the different metamorphic grade of these two units.
The most important ancient marble production site of the medium- to coarse-grained marbles of the Koralpe–Wölz System and of significance for this investigation are the marbles of Gummern. Decades ago, a series of authors reported on Roman mining traces there [14,15,16]. In the meantime, the ancient quarry faces and the remnants of the mining activities partly fell victim to the rapidly expanding modern mining industry. However, there are still minor signs of excavation in close vicinity of the modern quarry. The Roman quarries near Treffen occur in the same marble formation as the Gummern quarries, however, as will be shown below, they can be distinguished by multivariate discrimination analysis from the Gummern marbles. These marbles were frequently used in the region, and scientific data on the corresponding artefacts were already published [17,18]. The favourable logistic position on the banks of the Drau River was possibly the prerequisite of the export of these marbles to Southeastern Europe down the Danube [19].
The second marble-bearing formation of relevance in the considered area is bound to the Drauzug–Gurktal System. Also, for this group of marble quarries, a detailed description is given in Prochaska (2021), where six different quarries already used in Roman times were presented [12]. These formations are of greenschist metamorphic grade, and accordingly, the marbles of this group are fine-grained, varying from 0.1 mm to approx. 1.5 mm. A further characteristic of this group is an omnipresent silicate content (quartz, mica, chlorite, …). The quarries of Kraig, Seebichl, and Tentschach are located next to the Magdalensberg and are grouped together here as “Pörtschach I”. More distant are the locations Tiffen, Sekull, and Keutschach (“Pörtschach II”). As will be shown below, these two groups can easily be distinguished by their isotope characteristics.

3. The Investigated Samples of the Marbles from the Magdalensberg

Though major architectural elements and usable inventory were evidently transferred to Virunum, hundreds of fragments of marble artefacts remained on the mountain settlement. A collection of some gravestones and stele are on display on the site but are not addressed in this study. A large number of fragments from decorative architecture, like inscription plates, profiles, veneers, etc., were unearthed during the Magdalensberg excavations (Figure 2). Practically all of these objects are very small fragments of a few cm in size, and therefore, their original function is not always clear. Within a project aiming to catalogue the stone inventory, a set of 36 marble samples selected from the above-mentioned fragments was sampled, and the results are presented in this paper.
The artefacts were visually classified according to their macroscopic characteristics and classified into three different lithologic groups:
  • Group 1: Fine-grained to very-fine-grained samples (“Carrara type”)
This type has a white-to-slightly-greyish and dark-grey colour. Lustrous calcite crystals characterize the very white varieties, while the slightly greyish samples are extremely fine-grained with a dull and sandy surface (Carrara Bardiglio).
  • Group 2: Fine-grained samples of local origin (“Pörtschach type”)
Typically, the grain-size of this type is approx. 1 mm, and rare varieties with slightly larger grains occur. Muted shades of white, brownish, and slightly pink colours are dominant. The omnipresent but variable silicate content can usually be detected with the naked eye.
  • Group 3: Medium- to coarse-grained samples of local origin (“Gummern type”)
A grain-size of the calcite crystals between 2 and 5 mm is characteristic for this type. The colour is usually very white to slightly greyish. Visually, silicate accessory minerals can be very rarely detected.

4. The Applied Methods

  • For a detailed description of the analytical methods applied, reference is made to Prochaska and Attanasio 2021 [20]. An abbreviated summary of the methods used is given below. All analytical variables analysed are given in the supplementary materials.
  • Microscopy: If possible, an investigation of the artefacts with the petrographic microscope was carried out. Not all samples taken from the artefacts were big enough to prepare a thin section, however, examples of the microscopic textures of each group of artefacts will be presented. Furthermore the petrographic characteristics of the Bardiglio quarry samples will be discussed in Section 7.1.
  • Stable isotope analysis: The analysis of the stable isotopes of O and C is a standard method in marble provenance analysis [21,22]. For this investigation the samples were analysed in the laboratories of the University of Leoben/Austria.
  • Trace element analysis: The elements analysed were: Mg, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Pr, Dy, Ho, Yb, Pb, and U. The chemical analyses of all samples taken during this investigation were performed using ICP-MS analysis after dissolving the carbonate phase by hot HNO3. As this is a “near total analysis”, the trace element data obtained by this method must not be compared to data obtained by bulk analytical methods (e.g., HF dissolution, XRF analysis, etc.). Utmost care has to be exercised in sample preparation and handling in general as many significant trace elements occur only in the sub ppm range.
  • Fluid inclusion analysis: The extraction and analysis of the so-called “fluid-inclusions” in the calcite crystals provides significant proxies and additional variables for the characterization of marbles from different sources. The following components were analysed: DS, Li/Na, Cl/Na, K/Na, Br/Na, I/Na, and SO4/Na. The method was successfully applied in marble provenance analysis [20,23,24].
  • The statistical evaluation: The large number of variables requires multivariate statistical computing when a large number of variables are simultaneously computed. Of course any suitable statistical tool may be used to compute these data. The program packages STATISTICA and SPSS were used in this work.

5. The Analytical Results of the Investigated Samples

In the following, the analytical results for the different groups of marbles (according to Section 3) analysed for this work are given below. The analytical data of the analysed artefacts are given in Table S1. Numerical results of the corresponding quarry samples for comparison were published in Prochaska (2021) for the coarse-grained alpine marbles [12] and in Prochaska and Attanasio (2022) for the white Carrara marbles [13]. The database for the Carrara Bardiglio quarry samples used in this work is given in Table S2.

5.1. Analytical Results of the Fine-Grained (Group 1, “Carrara”) Samples

According to their visual appearance and their characteristic isotope numbers, 16 fine-grained or very-fine-grained samples were selected and grouped a priori as “Carrara” samples. All transitions do occur from very-white fine-grained Carrara samples to dark-grey almost-black very-fine-grained Bardiglio marbles. The analytical results of these samples were checked against the Carrara database samples [13] and against 52 Bardiglio quarry samples presented here in this work (Table S2). Out of this collection, 13 samples were selected due to their different petrographic and chemical characteristics (isotopes, Mn, Fe, REE, etc.) and are tentatively treated here as “Bardiglio II”. Clearly, the samples of the latter group comprise the darker varieties of the Bardiglio samples. The problems with the definition of grey marbles in general and the terms Bardiglio, Bigio, Nero di Colonnata, etc. will be discussed below.
The results of the stable isotope investigations are graphically displayed in Figure 3. The corresponding data fields are shown as statistical 90% ellipses. It is evident that the two types of Bardiglio can already be safely separated solely by isotope investigation. However, on this basis an almost complete overlap of Carrara white and Bardiglio exists. Except two outliers, all other samples investigated are within this overlapping area.
To enhance the discrimination of the different groups and to improve the separation of the data fields, further variables were included in the calculations.
The multivariate discrimination analysis computes a large number of analysed variables simultaneously and reduces them to a smaller number of artificial variables, so-called factors. For this evaluation procedure the variables δ18O‰, δ13C‰, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, and I/Na were used. As demonstrated in Figure 4, the data fields of the different types of quarry samples can be completely separated. Furthermore (except the two outliers) the majority of the samples plot in the Bardiglio data field and five samples are identified as Carrara white marble.
The two outliers observed already in the isotope diagram are still present in this multivariate evaluation. Compared to the remaining artefacts, in addition to the differences in the isotope composition, the two samples also show extremely high Sr contents.

5.2. Analytical Results of the Fine-Grained Marbles of Local Origin (“Pörtschach Type”)

The characteristics of the “Pörtschach type” marbles that were mined in several different sites in the Roman province of Noricum were previously reported in detail where the special features of these marbles are recorded in detail [12]. Compared to the similarly fine-grained Carrara marbles, the obvious general differences of the Alpine marbles dealt with here are the substantially higher concentrations of characteristic trace elements like Fe, or Mn of the latter. The aim of the procedure presented below is to try intra-site discrimination and to assign the artefacts of this group of the Pörtschach marbles to a corresponding single quarry.
The striking chemical feature typical for this type is the light δ18O isotope composition for all different marbles of the region (Figure 5). Despite the overlap of the compositional fields two major groups can be distinguished. The marbles of the Sekull, Tiffen, and Keutschach quarries exhibit slightly lighter isotope compositions and evidently do not coincide with the investigated artefacts. For the sake of clarity, these quarries are summarized as Pörtschach II in the following (Figure 6). In the isotope diagram the Pörtschach-type group of artefacts clearly falls in the overlapping field of the Kraig, Seebichl, and Tentschach quarries. However, solely on the basis of stable isotope analysis, the marbles of these quarries cannot be distinguished from each other.
To attribute these artefacts more precisely to their origin, and more specifically to a distinct quarry, a large number of variable were included in the evaluation. By the use of the analytical results of δ18O‰, δ13C‰, MgCO3, Mn, Fe, Sr, Cr, V, Y, Ba, La, Li/Na, and Cl/Na, a very good discrimination of the three quarry sites under consideration can be achieved. This procedure makes it possible to attribute five of the artefacts to the Seebichl quarry, with two additional artefacts being plot into the Kraig data field (Figure 6).

5.3. Analytical Results of the Medium- to Coarse-Grained Marbles of Local Origin (“Gummern Type”)

Among the total of 36 sample artefacts, a total of 13 turned out to be medium- to coarse-grained. The visual characteristics of the marble of these samples closely matched those of the marble from the Gummern quarry. This marble was widely used in Roman times in the area of Noricum and was also traded far beyond the provincial borders. A prerequisite for this was the favourable location of the quarry on the banks of the Drau River, which is a tributary of the River Danube. A detailed description of the situation of the Gummern marbles was published previously, and in the following, only condensed information is given [12]. A subgroup of the quarry samples from Gummern exhibited very unusual and extremely high Sr contents, and for the statistical analysis presented below, this subgroup is treated as “high Sr-Gummern”.
Traces of Roman mining at the Gummern quarry were largely removed due to intense mining activities during the last decades. However, some very rare toolmarks can be found in the vicinity of the present mine. The marble formation belongs to the “Wollanig” mountain range, extending for several kilometres. In the same geologic formation, some 10 km directly west, the Roman marble quarries of Treffen are very well preserved. As this marble belongs to the same formation as the Gummern marble, their characteristics are very much alike. Nonetheless, as will be shown below, they can to some extent be separated from the Gummern marble by factor analysis.
Another prominent medium- to coarse-grained marble in the province of Noricum already used in Roman times was mined at the Spitzelofen quarry [25]. The petrographic features are somewhat different to the investigated medium- to coarse-grained samples from the Magdalensberg. However, as the Spitzelofen quarry was evidently an important production centre, a possible provenance from this quarry was also tested and included in the statistical calculations presented below.
In the isotope diagram (Figure 7), the compositional fields of the selected quarries (Gummern, high Sr-Gummern, Treffen, and Spitzelofen) are superimposed upon each other to a large degree. Furthermore, the isotope composition of the Spitzelofen samples shows a very wide scatter and an accordingly very large 90% ellipse. Thus, a safe discrimination solely on the basis of isotope analysis is not possible for the investigated samples plot in the overlap area (except one outlier).
The essence of the discriminant analysis of the regional medium- to coarse-grained marbles is demonstrated in the bivariate plot of the canonical factors 1 and 2 (Figure 8). It shows that the majority of the projection points of the artefacts are within the fields of the different Gummern marbles. Five samples exhibit very high Sr-values, which are well known in the Gummern dataset of the quarry samples. Seven samples plot in the field of the “regular” Gummern dataset. One sample is on the interface between the Gummern and the Spitzelofen fields. However, the calculated probability tends towards Spitzelofen provenance for this sample.

6. The Petrographic Characteristics of the Different Groups of Artefacts

More detailed information on the petrographic characteristics of the fine-grained marbles from Carrara (white Carrara), the Pörtschach marbles, and the medium- to coarse-grained Gummern marbles are published elsewhere [12,13]. In the following, examples from each group of the investigated artefacts are presented. As no comprehensive chemical or petrographic data exist for the Bardiglio-type marbles, a description and discussion of these characteristics will be presented in Section 6.
  • Group 1 artefacts: Sample no. 878 (Figure 9a) is of the Carrara type and a typical representative of white Carrara marble. This sample is fine-grained with a maximum grain size well below 1 mm and exhibits a relatively regular homoeoblastic fabric with straight or slightly curved grain boundaries. These calcite marbles are usually very pure with very low amounts of accessory minerals that occur sporadically, including feldspar, quartz, opaque minerals, etc. In the present sample none of these trace minerals can be observed. One example of the “Bardiglio-type” of Carrara marbles is shown in Figure 9b (sample no. 895). It is very-fine-grained, and its greyish colour is due to fine, dispersed organic matter.
  • Group 2 artefacts: Sample no. 251 (Figure 9c) belongs to the local Pörtschach marbles. It exhibits a maximum grain size of about 1 mm. In accordance with the Pörtschach marbles in general, the marble of this sample is rather impure and accessory or trace minerals occur in the range of a few percent. The texture is characterized by a clear foliation due to the linear alignment of mica and chlorite. Further accessory minerals are quartz, apatite, and occasional graphitic pigment.
  • Group 3 artefacts: 13 medium- to coarse-grained artefacts that exhibit all petrographic characteristics of the local marbles from the Gummern quarries. The grade of deformation of these marbles may change across small distances, thus the textural features are highly variable. Figure 9d is an example of a moderately deformed Gummern marble (no. 394) with large calcite cristalloblasts and a moderate formation of smaller subgrains at the rims of the serrated calcites. Accessory minerals are quartz, mica, apatite, rarely tremolite, and graphite.

7. A Databank for the Carrara-Bardiglio Marbles

This chapter is intended to provide information on a very-fine-grained variety of Carrara marble usually called “Carrara Bardiglio”. Though databases exist on the regular white Carrara, no reference is made to the Bardiglio varieties, e.g., Gorgoni et al. [26]. The definition of different types of grey and black marbles is not very consistent in the literature and will not be discussed here. For a discussion about the problems in the nomenclature with grey and black marbles, studies such as Yavuz et al. and Attanasio et al. [27,28,29,30,31] can be referred to. In contrast to the great popularity of white Carrara marble, the Bardiglio varieties of Carrara marbles lead a shadowy existence in antiquity. Most probably, this marble was rather used for some architectural applications than for sculpture. However, a big modern industry based on these greyish Carrara varieties exists. To provide a tool to characterize the Bardiglio marbles by scientific means and to prevent a confusion of these marbles with the large number of fine-grained grey-to-black marbles used in antiquity, the relevant scientific parameters are presented below. The samples are from the collection of Donato Attanasio and are presently stored in the Austrian Archaeological Institute in Vienna.

7.1. The Petrographic Features

The term “Bardiglio” is used for fine- to very-fine-grained marbles from the Carrara basins with a colour ranging in a continuous transition from very light grey to almost black. The general observation is that the light varieties are of coarser grain compared to the dark and black marble, despite being of the same metamorphic grade.
The 52 Bardiglio quarry samples in our databank available for this study are from the sample collection of Donato Attanasio. As can be seen in Table S2 (supplementary online material), most of these samples were taken in the Colonnata basin. Because of their peculiar composition, 13 samples from Colonnata were separated and tentatively combined to the subgroup “Bardiglio II”. Certainly, there are singular samples with similar composition in other Carrara regions, but these are not included in the Bardiglio II group in this paper. Possibly, the “nero di Colonnata” also belongs to these very dark Bardiglio varieties.
White-to-light-grey varieties: With the unaided eye the light-grey Bardiglio varieties resemble very much the very-fine-grained varieties of the white Carrara samples. However, the Bardiglio samples are, in general, of even finer grain, in the range of 0.1 mm (Figure 10a,b). Under the petrographic microscope, the Bardiglio marbles exhibit a more or less pronounced schistosity. These light-grey varieties are calcite marbles with accessory feldspar, pyrite, apatite, and very subordinate amounts of organic matter. The texture is equigranular, but occasionally recrystallized lens layers do occur with calcite crystals up to 1 mm.
Dark-grey-to-black varieties: These varieties are distinctly of finer grain compared to the light samples. The schistosity is much more pronounced here due to layers of organic pigment and silicate impurities, like quartz and mica. Furthermore flakes of mica are aligned along these schistosity planes. Frequently occurring elongated lenses consisting of recrystallized calcite crystals and quartz are characteristic for the darker Bardiglio samples. It is estimated that the overall amount of non-carbonate minerals is in the range of a few percent. The dark colour of these samples is mainly due to the different amounts of organic matter along the schistosity planes, in the matrix, and, to a minor degree, to the presence of pyrite (Figure 10c–f).

7.2. The Geochemical Characteristics

One endeavour of this paper is to make a database for Carrara Bardiglio marble available in order to pinpoint the marbles of artefacts made of greyish, possibly-Bardiglio marble. As can be seen in the isotope diagram in Figure 3, the relatively small field of the Carrara white samples overlap almost completely with the light Bardiglio samples, which can, to some extent, be confused with white Carrara marbles. However, by using a larger set of variables evaluated by discrimination analysis, a very clear separation of these two varieties can be obtained. The evident difference is the slightly higher contents in Fe, Sr, Y, and the REE in the light-grey Bardiglio samples. The dark-grey-to-black Bardiglio samples are characterized by high O- and lower C-isotope values compared to the light Bardiglio and white Carrara samples. Furthermore, they exhibit much higher contents of most trace elements, especially Fe, Sr, Y, and the REE. It is evident that by using a wider range of variables (isotope, several trace elements, results of fluid inclusion analysis) a very good discrimination can be achieved.
As these dark Bardiglio varieties visually often resemble other grey-to-black marbles, these chemical features may help when discriminating dark-grey or black marbles. Therefore, the analytical data of the quarry samples are presented in Table S1.

8. Discussion and Conclusions

For a few decades, the emporium of the Magdalensberg was a trading hub in southern Noricum between the Roman Empire and the northern province. Iron ore deposits are rare in the Italic mainland, and accordingly, the supply had to be secured from other sources. Already from the 3rd century BC, findings of steel artefacts (swords) made most probably of Noric iron are reported [32]. The need for raw materials and especially the rising demand for high-quality steel for the armament of the Roman army within the course of the Gallic wars certainly triggered the upswing of the steel production in southern Noricum. This economic boom included all kinds of trade, and in this context, marble imports (as return freight?) into Noricum took place. These trading relations brought the “Roman way of life” to the northern provinces. Luxury marble decoration and architecture was essential for the self-image and identity of the Roman elites, and consequently a local marble production was established in Noricum. The first dated evidence is reported from findings of the Spitzelofen quarry from early imperial times [25]. A large number of quarries were opened in southern Noricum, and their marbles can be found all over the province. Of first importance was the marble from the Gummern quarries prominently present in the investigated Magdalensberg samples. Further examples of local marbles are the fine-grained Pörtschach marbles widely used on the Magdalensberg and in the area in general. For 16 artefact samples, a Carrara origin could be determined. It is worth mentioning that not only the white Carrara qualities were imported (5 samples) but also prominently the less valued and possibly cheaper Bardiglio marbles (9 samples). Two samples of this group are outliers and are most probably a variety of the grey Bardiglio marbles as they are extremely rich in Sr. Up to now, these Carrara artefacts are the only examples of imported marbles found on the Magdalensberg. This is in contrast to the findings in Virunum, which is the successor after the abandonment of the emporium and the capital of Roman Noricum, where a large number of sculptures where unearthed. Here, the spectrum of the marbles used is much broader, and in addition, all kinds of local marble, imports from Prokonnesos, Thasos, and Dokimeion can be found. The study of the provenance of the marbles used there is presently under investigation [33].
The number of grey-to-black marbles used throughout history is legion, and it is far beyond the scope of this paper to give adequate reference to this vast literature. To aggravate this situation, there is no precise definition of the different names for the black marbles, and the transition from limestone to marble in a petrographic sense is also blurring. Several occurrences are known, e.g., in the Mani peninsula, in Chios, Teos, Ephesos, Aphrodisias, or Tunisia and many more. Many of these deposits were studied in the past and a series of authors mention the use of Bardiglio in Roman times, evidently by visual examination but without proof or presentation of numerical analytical data [34,35,36,37,38,39,40]. In general, only scarce analytical data are available, and the data at hand are far away from being a comprehensive databank. A few authors present limited analytical data of a few locations of grey-to-black marbles used in antiquity [27,28,29,30]. For the efficient and successful discrimination of the various locations in any case, in a first step, a petrographic characterization has to be performed [41]. Thus, a pre-selection (e.g., for a statistical evaluation) can be obtained, which may help to avoid an overload of the calculations and the diagrams.
The analytical data of the quarry samples underlying this work are presented in the “Supplementary Online Material” (Table S2).

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/min13081036/s1, Table S1: The analytical data of the investigated artefacts; Table S2: The analytical data of the Carrara Bardiglio quarry samples.

Funding

This research received no external funding.

Data Availability Statement

The numerical data of the samples presented in this paper can be downloaded from the supplementary material in this paper. Furthermore data from ancient white marble quarries can be found under: https://doi.org/10.1016/j.jasrep.2022.103582; https://doi.org/10.1016/j.jasrep.2021.102958 (accessed on 25 July 2023) or upon request to the author.

Acknowledgments

I am very grateful to Vasiliki Anevlavi for the invaluable assistance with the laboratory work. My very special gratitude goes to my colleague and friend Donato Attanasio for many inspiring discussions from which I benefited so much. The Bardiglio quarry samples used in this paper are from his collection and are now stored in the Austrian Archaeological Institute in Vienna.

Conflicts of Interest

The author declares no conflict of interest.

References

  1. Prochaska, W. Die geologisch-lagerstättenkundliche Stellung der Hüttenberger Vererzungen. In The Production of Ferrum Noricum at the Hüttenberger Erzberg; Cech, B., Vienna, B., Eds.; Verlag Unipress: Graz, Austria, 2008; pp. 7–13. [Google Scholar]
  2. Cech, B. The production of ferrum Noricum. Largescale Roman iron production in Hüttenberg (Austria). In Acta Mineraria et Metallurgica. Studi in Onore die Marco Tizzoni; Cucine, C., Ed.; Notizie Archeologiche Bergomensi: Bergamo, Italy, 2012; Volume 20, pp. 91–106. [Google Scholar]
  3. Rehren, B.; Cech, B. Early Iron in Europe-an introduction and overview. Monogr. Instrumentum 2014, 50, 7–10. [Google Scholar]
  4. Doneus, M.; Gugl, C.; Jernej, R. Ein neu entdecktes römisches Militärlager in Virunum (Noricum). Erste Ergebnisse der Luftbildauswertung. Archäolog. Korrespondenzblatt 2003, 33, 393–402. [Google Scholar]
  5. Gugl, C.; Jernej, R. Virunum. Das Römische Amphitheater. In Die Grabungen 1998–2001; Hrsg, G.C., Jernej, R., Eds.; Wieser Verlag GmbH: Vienna, Austria, 2004. [Google Scholar]
  6. Thiedig, F.; Wolf, E. Römisches Bauen aus naturwissenschaftlicher Sicht in der Stadt auf dem Magdalensberg in Kärnten. Carinth. II 2003, 193, 33–128. [Google Scholar]
  7. Steiner, A. Auswertung Petrochemisch Untersuchter Römerzeitlicher Denkmäler aus Marmor in Kärnten; Magistrsko delo, Universität Wien: Wien, Austria, 2005. [Google Scholar]
  8. Steiner, A. Heimischer und importierter Marmor in Kärnten. Carinth. I 2006, 196, 67–78. [Google Scholar]
  9. Steiner, A. Südnorische Grabelemente und ihr Marmor. Frankf. Elektron. Rundsch. Zur Altertumskunde 2006, 1, 8–14. [Google Scholar] [CrossRef]
  10. Steiner, A. Die Funde aus Stein und Marmor auf dem Magdalensberg. Materialien. In Akten zum 2. Österreichischen Römerstein-Treffen 2006 in Salzburg, Hrsg.; Volker Höck; Felix Lang und Wolfgang Wohlmayr: Wien, Austria, 2007; pp. 123–125. [Google Scholar]
  11. Müller, H.W.; Schwaighofer, B. Die römischen Marmorsteinbrüche in Kärnten. Carinth. II 1999, 189, 549–572. [Google Scholar]
  12. Prochaska, W. The challenge of a successful discrimination of ancient marbles (part II): A databank for the Alpine marbles. J. Archaeol. Sci. Rep. 2021, 38, 102958. [Google Scholar] [CrossRef]
  13. Prochaska, W.; Attanasio, D. The challenge of a successful discrimination of ancient marbles (part III): A databank for Aphrodisias, Carrara, Dokimeion, Göktepe, Hymettos, Parian Lychnites and Pentelikon. J. Archaeol. Sci. Rep. 2022, 35, 103582. [Google Scholar] [CrossRef]
  14. Wolf, H. Die Kärntner Marmore und die Bausteine aus dem Leithakalke. In Wochenschrift des österreichischen Ingenieur- und Architekten-Vereins. 3; Verlag des Vereins: Wien, Austria, 1878; pp. 16–17. [Google Scholar]
  15. Jantsch, F. Der Römersteinbruch in Gummern. Carinth. I 1929, 119, 161–162. [Google Scholar]
  16. Kieslinger, A. Die nutzbaren Gesteine Kärntens. Carinth. II 1956, 17, 348. [Google Scholar]
  17. Djurić, B.; Hebert, B.; Hinker, C.H.; Hudeczek, E.; Müller, H.W.; Karl, S. Marmore römischer Brüche und Steindenkmäler in der Steiermark und in Štajerska. Ergeb. Eines Forschungsprojektes FÖ 2004, 43, 365–431. [Google Scholar]
  18. Djurić, B.; Müller, H.W. White Marbles in Noricum and Pannonia. An outline of the Roman Quarries and their Products. In Marbres et Autres roches de la Méditerranée Antique. Études Interdisciplinaires/Interdisciplinary Studies on Mediterranean Ancient Marble and Stones; Jockey, P., Ed.; ΛΕΥΚOΣ ΛΙΘOΣ: Paris, France, 2011; pp. 111–127. [Google Scholar]
  19. Djurić, B. The Logistics behind Ancient Art. The Case of Noricum and Pannoniae. Festvortrag. In Akten des 15. Internationalen Kolloquiums zum Provinzialrömischen Kunstschaffen. Der Stifter und sein Monument Gesellschaft–Ikonographie–Chronologie; von Steier, S., Porod, B., Scherrer, P., Eds.; Veröffentlichungen des Inst. für Archäologie der Karl-Franzens-Universität: Graz, Austria, 2019; Volume 16, pp. 8–30. [Google Scholar]
  20. Prochaska, W.; Attanasio, D. The challenge of a successful discrimination of ancient marbles (part I): A databank for the marbles from Paros, Prokonnesos, Heraklea/Miletos and Thasos. J. Archaeol. Sci. Rep. 2021, 35, 102676. [Google Scholar] [CrossRef]
  21. Craig, H.; Craig, V. Greek Marbles: Determination of Provenance by Isotopic Analysis. Science 1972, 176, 401–403. [Google Scholar] [CrossRef] [PubMed]
  22. Attanasio, D.; Brilli, M.; Ogle, N. The Isotopic Signatures of Classical Marbles; L´Erma di Bretschneider: Rome, Italy, 2006; p. 297. [Google Scholar]
  23. Prochaska, W.; Grillo, S.M. A new method for the determination of the provenance of white marbles by chemical analysis of inclusion fluids: The marbles of the mausoleum of Belevi/Turkey. Archaeom 2010, 52, 59–82. [Google Scholar] [CrossRef]
  24. Prochaska, W.; Attanasio, D. Tracing the origin of marbles by inclusion fluid chemistry. In Interdisciplinary studies on ancient stone: Proceedings of the IX Association for the Study of Marbles and Other Stones in Antiquity (ASMOSIA) Conference (Tarragona 2009); Institut Català d’Arqueologia Clàssica: Tarragona, Spain, 2012; pp. 230–237. [Google Scholar]
  25. Karl, S. Das römerzeitliche Marmorsteinbruchrevier Spitzelofen in Kärnten. In Montanarchäologische Forschungen; Beiträgen von, P.M., Bayer, M., Grabner, M., Hainzmann, R., Haubner, C.A., Hauzenberger, A.G., Heiss, K., Layr, D., Modl, W., Prochaska, S., et al., Eds.; Weißl und S. Wiesinger; Fundberichte aus Österreich Beihefte; Verlag Berger: Vienna, Austria, 2021; p. 1. [Google Scholar]
  26. Gorgoni, C.; Lazzarini, L.; Pallante, P.; Turi, B. An updated and detailed mineropetrographic and C-O stable isotopic reference database for the main Mediterranean marbles used in antiquity. In ASMOSIA V; Archetype Publications: London, UK, 2002; pp. 110–131. [Google Scholar]
  27. Yavuz, A.B.; Bruno, M.; Attanasio, D. An updated, multi-method database of Ephesos marbles, including white, greco scritto, and bigio varieties. Archaeom 2011, 53, 215–240. [Google Scholar] [CrossRef]
  28. Yavuz, A.B.; Bruno, M.; Attanasio, D. A new source of Bigio Antico: The ancient quarries of Iznik (Turkey). In ASMOSIA IX; Institut Català d’Arqueologia Clàssica: Tarragona, Spain, 2012; pp. 255–262. [Google Scholar]
  29. Attanasio, D.; Bruno, M.; Yavuz, A.B.; Prochaska, W. A multi-method database of the black and white marbles of Göktepe (Aphrodisias) including isotopic, EPR, trace and petrographic data. Archaeom 2015, 57, 217–245. [Google Scholar] [CrossRef]
  30. Attanasio, D.; Bruno, M.; Prochaska, W.; Yavuz, A.B. Ancient ‘Black’ Decorative Stones and the Ephesian Origin of Sculptural bigio antico. Archaeometry 2017, 59, 794–814. [Google Scholar] [CrossRef]
  31. Attanasio, D.; Bruno, M.; Prochaska, W.; Yavuz, A.B. Göktepe Marbles, White, Black and Two-Tone; L’Erma di Bretschneider: Rome, Italy, 2021; p. 323. [Google Scholar]
  32. Buchwald, V.F. Iron and Steel in Ancient Times. Kgl; Danske Videnskabernes Selskab: Copenhagen, Denmark, 2005; p. 372. [Google Scholar]
  33. Ladstätter, S.; Anevlavi, V.; Prochaska, W. Not always a question of distance! Marble provenance studies on Roman statues in Noricum. In Proceedings of the Presentation at the ASMOSIA XIII Conference, Vienna, Austria, 19–24 September 2022; p. 85. [Google Scholar]
  34. Pensabene, P.; Lazzarini, L. Il problema del bigio antico e del bigio morato: Contributo allo studio delle cave di Teos e Chios. In Marmi Antichi II; Pensabene, P., Ed.; Publisher L’Erma di Bretschneider: Rome, Italy, 1998; pp. 141–173. [Google Scholar]
  35. Bruno, M.; Pallante, P. The «lapis taenarius» quarries of Cape Tainaron (Mani Peninsula, S. Peloponnesus, Greece). In Proceedings of the ASMOSIA VI, Venice, Italy, 15–18 June 2000; pp. 163–176. [Google Scholar]
  36. Lazzarini, L.; Agus, M.; Cara, S. The ancient quarries of the neri antichi (black limestones) from Zeugitania (Tunisia). Marmora 2006, 2, 59–70. [Google Scholar]
  37. Agus, M.; Cara, S.; Lazzarini, L.; Mola, M. A laboratory characterisation of black limestones (neri antichi) from Zeugitania (Tunisia). Marmora 2006, 2, 71–82. [Google Scholar]
  38. Brilli, M.; Giustini, F.; Kadioğlu, M. Black limestone used in antiquity: Recognizing the limestone of Teos. Archaeometry 2019, 61, 282–295. [Google Scholar] [CrossRef]
  39. Taelman, D.; Antonelli, F. Provenance of the white and polychrome marbles used for the architecture and sculpture of roman Sentinum (Sassoferrato, Marche, Italy). Archaeometry 2021, 64, 1–24. [Google Scholar] [CrossRef]
  40. Taelman, D.; Mei, O.; Antonelli, F. On the Use of White Marbles in Roman Forum Sempronii (Fossombrone, Marche, Italy): An Overview of the Archaeometric Data on Architectural Elements. Heritage 2023, 6, 2942–2955. [Google Scholar] [CrossRef]
  41. Lapuente, P.; Nogales, T.; Leon, P.; Royo, H.; Preite-Martinez, M. Black Sculptural Materials from Villa Adriana: Study Provenance. In ASMOSIA IX; Institut Català d’Arqueologia Clàssica: Tarragona, Spain, 2012; pp. 376–383. [Google Scholar]
Figure 1. The locations of the different quarries in Noricum already used in Roman times and the location of the emporium of the Magdalensberg (modified after [12]). The dashed line is the approximate border between Noricum Ripense (in the north) and Noricum Mediterranum (in the south).
Figure 1. The locations of the different quarries in Noricum already used in Roman times and the location of the emporium of the Magdalensberg (modified after [12]). The dashed line is the approximate border between Noricum Ripense (in the north) and Noricum Mediterranum (in the south).
Minerals 13 01036 g001
Figure 2. Examples of fragments of the different groups of sampled artefacts from the Magdalensberg: (a) white Carrara; (b) Bardiglio; (c) Pörtschach type; (d) Gummern type.
Figure 2. Examples of fragments of the different groups of sampled artefacts from the Magdalensberg: (a) white Carrara; (b) Bardiglio; (c) Pörtschach type; (d) Gummern type.
Minerals 13 01036 g002
Figure 3. On display in the isotope diagram are the investigated artefact samples of the group “Carrara”. Also shown are the data fields (calculated 90% ellipses) of different Carrara quarry samples used for comparison (small dots, also in next Figures).
Figure 3. On display in the isotope diagram are the investigated artefact samples of the group “Carrara”. Also shown are the data fields (calculated 90% ellipses) of different Carrara quarry samples used for comparison (small dots, also in next Figures).
Minerals 13 01036 g003
Figure 4. Bivariate plot of the discriminant factors 1 and 2 using the following variables: δ18O‰, δ13C‰, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, and I/Na. The data fields are the same as in Figure 3. By using this combination, a complete separation of the data fields was achieved.
Figure 4. Bivariate plot of the discriminant factors 1 and 2 using the following variables: δ18O‰, δ13C‰, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, and I/Na. The data fields are the same as in Figure 3. By using this combination, a complete separation of the data fields was achieved.
Minerals 13 01036 g004
Figure 5. In the isotope diagram of the regional fine-grained marbles the compositional fields of the different quarries show considerable overlap. Two different major groups are indicated (see text).
Figure 5. In the isotope diagram of the regional fine-grained marbles the compositional fields of the different quarries show considerable overlap. Two different major groups are indicated (see text).
Minerals 13 01036 g005
Figure 6. Bivariate plot of the discriminant factors 1 and 2 using the following variables: δ18O‰, δ13C ‰, MgCO3, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, and I/Na. By using this combination, a very good separation of the data fields was achieved.
Figure 6. Bivariate plot of the discriminant factors 1 and 2 using the following variables: δ18O‰, δ13C ‰, MgCO3, Mn, Fe, Sr, Cr, V, Y, Cd, Ba, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, and I/Na. By using this combination, a very good separation of the data fields was achieved.
Minerals 13 01036 g006
Figure 7. Isotope diagram of the regional medium- to coarse-grained marbles under consideration. The compositional fields of the different quarries show considerable overlap.
Figure 7. Isotope diagram of the regional medium- to coarse-grained marbles under consideration. The compositional fields of the different quarries show considerable overlap.
Minerals 13 01036 g007
Figure 8. The first two canonical factors of the calculations for the regional medium- to coarse-grained marbles reveal a very good separation of the different source areas. The following variables used were: δ18O‰, δ13C‰, Mn, Fe, Sr, V, Y, Cd, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, Br/Na, and I/Na.
Figure 8. The first two canonical factors of the calculations for the regional medium- to coarse-grained marbles reveal a very good separation of the different source areas. The following variables used were: δ18O‰, δ13C‰, Mn, Fe, Sr, V, Y, Cd, La, Ce, Yb, DS, Li/Na, Cl/Na, K/Na, Br/Na, and I/Na.
Minerals 13 01036 g008
Figure 9. Microphotos under polarized light of the three different groups of the investigated artefacts. (a) (no. 878) is white Carrara, (b) (no. 895) is Bardiglio, (c) (no. 251) is Pörtschach-type and (d) (no. 394) is Gummern marble. The length of the images is 6 mm in all cases.
Figure 9. Microphotos under polarized light of the three different groups of the investigated artefacts. (a) (no. 878) is white Carrara, (b) (no. 895) is Bardiglio, (c) (no. 251) is Pörtschach-type and (d) (no. 394) is Gummern marble. The length of the images is 6 mm in all cases.
Minerals 13 01036 g009
Figure 10. Microphotos of different varieties of Carrara Bardiglio. (a,b) (no. 4393) are images of a very light variety, (cf) (no. 4383) represent a dark-grey type of Bardiglio II. The length of the images is 6 mm (ac) and 1.2 mm in (df).
Figure 10. Microphotos of different varieties of Carrara Bardiglio. (a,b) (no. 4393) are images of a very light variety, (cf) (no. 4383) represent a dark-grey type of Bardiglio II. The length of the images is 6 mm (ac) and 1.2 mm in (df).
Minerals 13 01036 g010
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Prochaska, W. Trading Marble for Steel: Early Roman Import of Carrara Marble into the Alps—The Example of the Magdalensberg Trading Post in Noricum. Minerals 2023, 13, 1036. https://doi.org/10.3390/min13081036

AMA Style

Prochaska W. Trading Marble for Steel: Early Roman Import of Carrara Marble into the Alps—The Example of the Magdalensberg Trading Post in Noricum. Minerals. 2023; 13(8):1036. https://doi.org/10.3390/min13081036

Chicago/Turabian Style

Prochaska, Walter. 2023. "Trading Marble for Steel: Early Roman Import of Carrara Marble into the Alps—The Example of the Magdalensberg Trading Post in Noricum" Minerals 13, no. 8: 1036. https://doi.org/10.3390/min13081036

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop