Vibe of Wildness and Death: A Multidisciplinary Study of the Arena Wall Decoration of the Amphitheater in Viminacium (Kostolac, Serbia)

Abstract

This paper focuses on the study of the arena wall decoration in the amphitheater at the archaeological site of Viminacium. The architectural characteristics of the amphitheater, along with the spectacle iconography, have made this finding one of the most interesting discoveries at Viminacium, as well as in a wider context. A multidisciplinary approach that included an iconographic and archaeological study, as well as Energy Dispersive X-ray Fluorescence (EDXRF), X-ray Powder Diffraction (XRD), and Raman and Fourier-transform Infrared (FTIR) spectroscopy analysis, was applied to determine the palette of the pigments used for the arena wall decoration and understand the iconography and its context in more detail. Among the commonly used earth pigments (yellow, red, brown, and green colors), copper-based pigments (green and blue Egyptian blue), and the most precious ones for the period—namely, cinnabar and lapis lazuli—were identified. The applied analytical techniques enabled a tentative suggestion of the origin of the raw materials of some of the pigments that were used, such as marine sediments or rocks from different destinations. Due to the fact that the Viminacium amphitheater constitutes a typical example of a provincial building reserved for public spectacles, the results of this study will significantly contribute to our understanding of the function of the amphitheaters in the Danubian region, as well as throughout the Roman world.

1. Introduction

Amphitheaters represent an essential symbol of Roman civilization. The earliest amphitheaters were built in southern Italy, and, as Rome expanded, they continued to be built in various areas, mostly in urban centers and near military forts [1,2,3]. Besides gladiatorial fights (munera gladiatorum), various shows with wild animals (venationes), which featured animal slayings, as well as executions of convicts (damnatio ad bestias), were held in them. The remains of amphitheaters are comparatively rare in Danubian provinces [4], and the amphitheater of Viminacium is the only securely attested within the territory of modern-day Serbia (today, a wider area of Selo Kostolac) [5]. The amphitheater of Viminacium stood out from other similar buildings by its painted decoration of the arena wall, which is unique in Roman art and bears direct symbolism of triumph in hunting or in combat with animals [6,7]. It consisted of two zones. The lower zone was decorated with a simplified imitation of marble paneling, while the upper zone consisted of bordures with depictions of stretched hides of large felines. When we think about this iconographic approach, it inevitably invokes images of fighters in the arena locked in brutal combat with lions, tigers, and leopards, while the audience watched the spectacle with alternating feelings of suspense, excitement, and astonishment.

Energy Dispersive X-ray Fluorescence (EDXRF), X-ray Powder Diffraction (XRD), and Raman and Fourier-transform Infrared (FTIR) spectroscopy analytical techniques were applied in this study to determine the palette of pigments used during the painting of the arena wall.

2. Archaeological Context

Viminacium is located in eastern Serbia, near the confluence of the rivers Mlava and Danube (Figure 1). This place was originally the location of a legionary fortress, where Legio VII Claudia was stationed from the second half of the 1st century onward. Viminacium developed as the capital of the province of Upper Moesia, next to the military fort. As a military, civil, economic, and cultural hub, Viminacium represents one of the most important ancient sites in modern-day Serbia. It is known for its wall paintings, which are primarily found in its late antique tombs [8,9,10], as well as in other buildings [6,11]. One of its most representative examples is the wall decoration of the amphitheater.

Systematic excavations of the amphitheater were undertaken between 2007 and 2017 (Figure 2 and Figure 3). In the initial phase of its construction, the amphitheater was made of wood and constructed next to the legionary fortress at the beginning of the 2nd century. During the first half of the 2nd century, this structure was replaced by one made of both wood and stone. This structure became a part of the city’s area by the end of the 2nd century and the beginning of the 3rd century, as the city expanded into its northeastern corner. The amphitheater was used until the first half of the 4th century. The building was abandoned at that time, and by the end of the 4th century, a necropolis was formed above its remains [5,7,12].

During the archaeological excavations in 2009, 2011, and 2013, collapsed fragments of wall paintings were discovered in the south and southwestern part of the amphitheater (sq. D/7, E/7–8, and F-G/8) next to the arena wall (Figure 4) [12,13,14]. Fragments of wall paintings were also discovered in 2011 in the northwestern part of the arena (sq. E/4), within a late Roman pit. The arena wall itself contained a couple of fragments of wall decoration, which indicates that these fragments belonged to the lower part of the painted unit, that is, the part that was leaning against the very level of the arena.

The arena wall was erected during the second phase of the amphitheater’s construction (Figure 3, Figure 5 and Figure 6). This phase is dated to the period from the first half of the 2nd to the first half of the 4th century [5,12]. It includes different stages that involved various modifications on wooden construction and minor adaptations on the walls, although the basis of the amphitheater and its overall dimensions remained unchanged. The arena wall represented a barrier between the participants and the spectators, functioning as a barrier between the audience and the arena stage. Its average height was 2.63 m [1]. Based on its construction techniques, two distinct phases of construction of the Viminacium amphitheater arena wall can be clearly distinguished (Figure 5).

Initially, the wall was built using the opus quadratum technique. It was made of limestone blocks bound with mortar and was preserved up to a height of 2.4 m. During the second phase, it was reconstructed in the opus incertum technique, using crushed shale and mortar, and its height was increased. Within the arena wall, on the longer axis of the building, there are monumental entrances into the amphitheater (porta sanavivaria and porta libitinensis). Besides these, five service doors (porta postica) were discovered, while it is assumed that there were two more passages that connected the arena to the smaller rooms behind the wall and below the stands. The analysis of the architectural elements indicates that the arena wall was as tall as 3.40 m above the arena in the final phase of its use (Figure 6) [5,12].

The height of the wall was determined according to the specific activities in the arena. In her PhD thesis, S. Vuković [15] argues that the provincial amphitheaters mostly used local animals. During the excavations of Viminacium, remains of a wild boar and brown bear were discovered [7,15,16]. However, the discovery of skeletal remains of a leopard in the wider area of the amphitheater [7] indicates that large felines also participated in the shows organized in the Viminacium arena. The presence of a brown bear and exotic animals in these spectacles certainly justifies the height of the wall of the podium in the Viminacium arena. The length of a bear could be up to 2.8 m and 1.25–1.5 m; thus, standing on their hind legs, these animals can even be above 3 m tall [17,18]. E. Dyggve concludes that lions can leap up to 2 m [19], while other data indicate that large felines can jump above 4 m in height [20]. On the upper edge of the arena wall, there was a thinner wall or a fence (balteus, pluteus) to increase its height and protect the spectators more efficiently (Figure 6). The height of this barrier ranged from 0.3 to 0.8 m; in certain amphitheaters, it was as high as 1–1.2 m [21,22,23]. The protection of the audience could have been enhanced by the addition of other elements, which were not defined in Viminacium, such as protective nets (cancelli) fixed to the top of the wall, but also additional fences, fitted in front of the arena wall itself [1]. The area intended for combat was often limited to the central part of the arena, so as to avoid a blind spot caused by the height of the arena wall [1] and to enable the audience to watch the games without hindrance.

Wall decoration was identified in smaller areas in situ, with no traces of corrections to the painting. Collapsed parts of the decoration were found right next to the arena wall, covered by an archaeological layer dated to the period between the middle and the end of the 4th century, the period when the amphitheater fell out of use. The arrangement of the fragments was irregular, with layers of fragments found overlapping, facing up or down, or positioned diagonally. Together with the findings of fragments of pictorial decoration in the pits from the mid-4th century, these data indicate that the decoration of the arena wall in question belonged to the final stages of the construction of the building. They are dated between the middle of the 2nd century and the first half of the 4th century, when, in all likelihood, the final adaptations of the wall in the southwestern part of the building were carried out (Figure 5).

3. Decoration of the Arena Wall

The iconographic representation on the arena wall is authentic. It consists of two zones: an upper and a lower one. The field separating the two zones is a black line, 2 to 3 cm wide, while above it, there is a painted red bordure, approximately 10 cm wide. According to the undertaken conservation and restoration works and the study of the wall decoration, it was determined that there were originally two zones of wall painting: an upper and a lower one. Unfortunately, the data about how this composition ended in the upper part is missing, so it remains unknown whether there was any painting above the upper zone or whether it simply ended with the bordure. In the fragments of the lower zone, whose height was probably about 1.5 m, lasure areas of red, green, and ochre were visible (Figure 7A). Afterward, it was deduced that it was a very simplified example of marble imitation. In the upper zone, three motive schemes in the form of a frame with decoration were defined. The first pattern was a black interweaving on an orange-red background (Figure 7B). The second pattern consisted of a repetitive motif in black-brown, resembling a U-shape design (Figure 7C1,C2). This motif was rendered on a background consisting of segments of various color blends: olive green, gray, and purple. Since 2018, when it was first examined, the iconographic imagery consisted only of these two patterns (Figure 8) [10]; another unit was conserved, and the third pattern was defined (Figure 7D1,D2 and Figure 9). It was rendered on an ochre background with hues of brown in the middle of a composition. These three patterns represent three distinct fields, separated with bordures: a red bordure flanked by yellow bands, each of them outlined with a white line. The analysis of the painted decoration determined that there were three alternating patterns on the arena wall of the amphitheater in Viminacium. They depicted stretched hides of a leopard, a tiger, and a lion (Figure 8, Figure 9 and Figure 10), and the approximate dimensions of the painted fields were about 1.5 × 1.5 m, but it is worth considering that not all the frames had identical dimensions.

The leopard’s “fur” was recognized first (second scheme), since the same motif is depicted on the cloaks of the participants in Dionysus’ procession on the mosaic “Triumph of Bacchus” from Sousse, Tunisia (central part, the scene framed with vines), where Bacchus himself is wearing a leopard skin dress, while a leopard drinking water from a kantharos is in the foreground (Figure 11) [24]. The same mosaic shows harnessed tigers with a motif very similar to the one from our first scheme. The mosaic from Sousse completely dispelled any doubts that there might have been any other motifs on the arched wall of the Viminacium amphitheater. The fur motif resembling our second scheme, only in a different color palette, can be seen on the feline from the wall painting in Villa of Poppaea (area 13) in Naples [25], while similarly painted fur can be seen on a fresco depicting a leopard from Villa of Cicero in Pompeii (now in Naples Archaeological Museum, Inv. Nu. s. n. Ant. Inv. 72).

Animals were painted in various Roman buildings, but depictions of stretched animal hides are extremely rare. On the peristyle of a villa from Yvonand (Switzerland) from the 2nd century, there are painted animal skins with representations of venators (Figure 12) [26]. There are more examples besides those from Yvonand and Viminacium; for example, the one from Komárom-Szőny (Brigetio), dated to the 3rd century, shows a depiction of stretched animal hides. In it, lion and panther hides can be recognized (Figure 13) [27]. In addition to this, an example from Villa of Els Munts (Altafulla, Tarragona) needs to be mentioned [28,29,30].

Bizarre as it may seem, the imitation of stretched hides represented triumph or a trophy, which is completely understandable in the context of a spectacle in an amphitheater. It should also be taken into account that these wild cats were brought to Viminacium from remote parts of the Empire; thus, such a decoration conveyed an exceptionally exotic trait [10].

The very function of the arena wall required it to be painted in simple motifs, especially in the lower zone of the wall. In some cases, the entire wall was covered in marble paneling, which, among other things, made it difficult for the animals to climb up [1,3]. The painted decoration required painting the wall in one color, most often red. A similar simple decoration of the arena wall was discovered in the military amphitheater in Chester (Deva Victrix) [31], as well as in the civilian amphitheaters in Petronell (Carnuntum), Budapest (Aquincum), and Solin (Salona), but also in other buildings across the state [1]. The examples that depict the very scenes from the arena are rare. In the Pompeian amphitheater, the painted decoration included depictions of gladiators and scenes with animals, that is, hunting and animal combat [32]. Segments of a painting portraying animal combat and hunting scenes were discovered in the amphitheater in Mérida (Augusta Emerita) [33,34]. The scenes of hunting were also depicted on arena walls of buildings in Larache (Lixus) [1,35] and Corinth (Colonia Corinthiensis) [1], but also on the wall of the podium of Herod’s Circus in Caesarea (Caesarea Maritima) [20,36], which was a multifunctional building and was adjusted to this type of spectacle. Just like in the lower zone of the amphitheater in Viminacium, the painted decoration of the arena wall could have imitated architectural elements. On the fresco from Pompeii, which shows an amphitheater and a fight between the residents of Pompeii and Nocera, it was noted that the arena wall was decorated in such a way as to mimic marble plates [32]. In the semi-amphitheater in Augst (Augusta Raurica), the lower zone of the arena wall comprised decorative fields, with depictions of polychrome marble, while above them, there were floral motifs on a white background [22,37,38]. The arena wall of the amphitheater in Augusta Raurica had polychrome decoration, which represented an imitation of marble plates [22,38]. It was established that the painted decoration in Cirencester (Corinium Dobunnorum) imitated the appearance of marble [39].

4. Materials and Methods

Fresco fragments were removed from the site during several conservation campaigns from 2010 to 2013. To identify the pigment palette, six fragments of wall decorations were chosen for analytical examination. As the art scheme repeats itself, it was not necessary to obtain a large number of samples. Samples of a thicker layer of a particular color with sufficiently large areas of clear nuance were chosen. The selected fragments are listed in

Table 1. Arena wall decoration fragments used in this study.

ID	Fragment Photo	XRF	XRD	Raman	rFTIR
F1		3 red	x	x	x
		3 ochres	1 ochre	x	x
		1 white	x	x	x
F2		3 dark red	1 dark red	x	x
F3		1 green	1 green	x	x
F4		x	1 grayish blue	1 grayish blue	1 grayish blue
KV/E8		1 red	1 red	x	x
		1 black	x	1 black	x
KV/E7-E8		1 red	1 red	x	x
		1 yellow	1 yellow	x	x
		1 green	1 green	x	x
		1 blue	x	x	x
		1 white	x	x	x

.

Some of the pigments can be identified according to their elemental composition. For this purpose, the non-invasive, non-destructive, well-established, and commonly used pEDXRF spectrometry method was applied. The measurements were performed using a pEDXRF spectrometer, constructed at the Vinča Institute of Nuclear Sciences in Belgrade, consisting of an air-cooled X-ray tube with a pin-hole collimator (Oxford Instruments, Rh-anode, maximum 50 kV, 1 mA). A SiPIN X-ray detector (6 mm²/500 μm, Be window with 0.5 mil/12.5 μm thickness and a 1.5-inch detector extension) along with a DSP (X123, Amptek Inc.) for spectra acquisition was also integrated into the instrument. The precise positioning and visualization of the measurement spot were ensured by two laser pointers. The following experimental parameters were maintained the same during all measurements: no filter, 40 kV high voltage, 800 μA current, 100 s measuring time, distances between the sample and the X-ray tube, and the sample and the detector of 21 mm and 22 mm, respectively. The spectra acquisition and elemental identification were conducted using ADMCA software (Amptek Inc., version 1, 0, 0, 16).

For more precise pigment identification, additional XRD analysis was conducted. The samples were analyzed on an automatic powder diffractometer Rigaku SmartLab under the following conditions: copper anticathode radiation with a wavelength of CuKα = 1.54178 Å, operating tube voltage U = 40 kV, and current I = 30 mA. The samples were analyzed in the range of 3–75° 2θ with a step of 0.01° and a data collection rate of 5.0°/min. Based on the obtained I/Imax values and inter-planar spacings d, by comparing them with the literature data and ICDD PDF-2 (2016) standards, the present crystalline phases (minerals) were identified.

Since the identification of black and greyish blue pigments was not possible using the above-mentioned analytical techniques, the KV/E8 and F4 fragments were analyzed by applying Raman spectroscopy (Gem Raman System TO-GRS-532/TO-GRS-532-E—Thunder Optics). The FTIR spectra at the specified position on fragment F4 were obtained using a Bruker Optics ALPHA-R portable spectrophotometer (Bruker Optik GmbH in Ettlingen, Germany). The measurements were conducted at room temperature utilizing an external reflection module with optics (22°/22°), with a spot size of approximately 4 mm in diameter. The analysis covered a range from 400 to 4000 cm⁻¹, averaging 64 scans at a resolution of 4 cm⁻¹.

5. Results and Discussion

Red areas were analyzed by EDXRF spectrometry on the fragments F1 (three different spots), KV/E8 (one spot), and KV/E7-E8 (one spot). The comparative EDXRF spectra of three different spots analyzed along the red area on fragment F1 are presented in Figure 14a. It was determined that cinnabar was used as a coloring pigment, together with some earth pigments. Cinnabar was used uniformly along the fragment area, while an iron-based pigment (most probably red earth and/or umber) was added to reach the desired shade or to reduce the usage of precious cinnabar. The red areas were analyzed as two more fragments (KV/E8 and KV/E7-E8). According to the EDXRF spectra (Figure 14b), red-colored earth pigments were detected in both fragments. Additional XRD and Raman spectrometry analysis confirmed the usage of a hematite-based red earth pigment (Figure 15a).

The three different dark red to brownish colored spots were analyzed on fragment F2 by EDXRF spectrometry. A more or less homogenously spread pigment was identified as an earth pigment with the highest content of manganese among all analyzed samples (Figure 14c). According to the XRD analysis, which identified pyrolusite and illite minerals (Figure 15b), the pigment in question may be clay-rich umber [40,41].

Yellow areas were present on two fragments: F1 and KV/E7-E8. Using EDXRF spectrometry, the usage of an earth pigment was discovered by the identification of iron with impurities of K, Ti, Mn, and Sr (Figure 14d). Additional XRD analysis (Figure 15c) revealed the usage of the yellow ochre pigment, by the presence of the goethite mineral.

The EDXRF analysis of the green areas on two fragments (F3 and KV/E7-E8) indicated the use of the green earth pigment, which was applied singly, without mixing with other pigments (Figure 14e). The mineral composition of the green pigments indicated the use of green earth from two different sources of raw material [42]. On one fragment, the green earth was characterized solely by the presence of glauconite (fragment F3, Figure 15d), indicating sedimentary rocks of marine origin (green sand) [43]. The green pigment composition on fragment KV/E7-E8, besides glauconite, contained celadonite, a mineral also characteristic of green earth, which tentatively indicates the use of a raw material originating from Cyprus or Italy (Figure 15e) [42].

The blue pigment was present only on the fragment KV/E7-E8 and was analyzed only by EDXRF spectrometry since the usage of Egyptian blue was confirmed (Figure 14f).

The pigment palette also included chalk white (Figure 14g) and vine black. The vine black pigment was identified by Raman spectrometry (Figure 16a) by comparing the fragment’s spectrum with the database spectra for black pigments. By the XRD analysis, the sodalite mineral was identified in the grayish blue parts of the fragment F4 (Figure 15f).

This result indicated the use of sodalite-group mineral pigments, like lapis lazuli or, most probably, the ultramarine ash pigment (the cheaper version of the blue lapis lazuli pigment). The XRD finding was confirmed by Raman (Figure 16b) and rFTIR spectroscopy (Figure 16c). By identifying the characteristic of a weak band at 583–587 cm⁻¹, which correspond to antisymmetric stretching vibrations of the S₃⁻ radical anion, the ν₃ mode, and a lower intensity of the band at 1139 cm⁻¹ (associated to S−O stretching vibrations), together with a narrow and weak band at 2340–2342 cm⁻¹ (due to antisymmetric stretching vibrations of CO₂ molecules) and 2275 cm⁻¹ (a very weak isotope satellite corresponding to ¹³C¹⁶O₂), the use of lapis lazuli was confirmed [44].

The summarized data on the identified pigments are provided in Table S1 in the Supplementary Materials.

6. Conclusions

The wall decoration in the Viminacium amphitheater represents a unique case, thus far, when compared to other Roman amphitheaters. In Roman iconography, events from the arena include various depictions, such as gladiatorial combats, hunts, and combat with animals. Depictions of large felines, namely, leopards, tigers, and lions, constitute an integral part of the known iconography, but not a single scene of stretched hides of these animals has been hitherto recorded in Roman wall paintings.

The decoration of the arena wall itself complemented the atmosphere in which the games were held, clearly alluding to the events staged in the amphitheater—particularly the elaborate animal shows. The representations of stretched animal hides, together with the leopard bones found in the wider area of the amphitheater, testify to the fact that, besides local wild fauna, large felines were imported from remote areas for the games in Viminacium. The representations of leopard, lion, and tiger furs on the arena wall symbolize the trophies awarded to the participants who successfully slayed the animals in the arena.

An incredibly rich palette of pigments was used for the wall decoration. Besides the commonly used red, yellow, and green earth pigments, the most commonly used blue pigment, Egyptian blue, was also detected. Vine black and chalk were also identified. Earth pigments were used individually, but the red earth pigment was mixed with cinnabar, most likely to reduce the use of the more precious pigment. The mineralogical analysis revealed the possible origin of the pigments’ raw materials. The so-called green sand pigment was determined by identifying solely glauconite in the green earth pigment. The presence of celadonite in addition to glauconite in the green earth pigment indicates the Cyprus or Italian origin of raw materials for the pigment. The detection of the sodalite mineral was an indication of the usage of a very expensive lapis lazuli pigment. The discovery that pigments from distant regions, together with the precious cinnabar and lapis lazuli pigments, were used for the arena wall decoration was unexpected yet not uncommon, given the discoveries of animal remains from afar. The unique iconographic scheme, together with the pigments and the wild animals brought from distant parts of the Empire, indicates the cosmopolitan character of Viminacium. The use of the amphitheater coincided with a period of prosperity in Viminacium. During this period, a variety of goods, including pigments, were imported from across the Roman world. Simultaneously, a local painting officina was active, which helps explain the incredibly rich palette of pigments used to decorate the arena wall of the Viminacium amphitheater.

The specificity of the decoration itself reflects the importance of the amphitheater for the residents of Viminacium. Holding gladiator fights in the arena was not only entertainment for the crowds, but also a means of promoting the Roman way of life and customs. The Imperial cult was extolled in the amphitheaters, so decorating the arena walls may have been intended to enhance the atmosphere during visits by several Emperors, such as Hadrianus, Septimius Severus, Caracalla, Gordianus III, Diocletianus, and Constantinus, who resided in Viminacium. Wall paintings may have been commissioned to honor other important occasions, such as the proclamation of Viminacium’s elevation to the status of a colony, various public or civic holidays, or celebrations of military triumphs by the Roman army in campaigns involving members of the VII Claudia legion.