Next Article in Journal
Identification of Alteration Minerals and Lithium-Bearing Pegmatite Deposits Using Remote Sensing Satellite Data in Dahongliutan Area, Western Kunlun, NW China
Previous Article in Journal
Mineralogy and Fluid Inclusion Constraints on the Genesis of the Recently Discovered Ag-(Ni-Co-Sb-As-Hg ± Bi) Vein Ore Shoot Mineralization in the Aouli Pb-Zn District (Upper Moulouya, Morocco)
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

White Marble Sourcing and Regional Workshop Dynamics in Roman Thrace: An Archaeometric Study of Votive Reliefs

1
Austrian Archaeological Institute, Austrian Academy of Sciences, Dominikanerbastei 16, 1010 Vienna, Austria
2
National Archaeological Institute with Museum, Bulgarian Academy of Sciences, 2 Saborna Str., 1000 Sofia, Bulgaria
3
Department of Classical Archaeology, University of Vienna, Franz-Klein-Gasse 1, 1190 Vienna, Austria
*
Author to whom correspondence should be addressed.
Minerals 2025, 15(7), 670; https://doi.org/10.3390/min15070670 (registering DOI)
Submission received: 25 April 2025 / Revised: 31 May 2025 / Accepted: 11 June 2025 / Published: 22 June 2025
(This article belongs to the Section Mineral Geochemistry and Geochronology)

Abstract

:
Marble votive reliefs from rural sanctuary contexts have seldom undergone archaeometric investigation, despite their potential to illuminate regional trade and production networks. This study focuses on such reliefs from Philippopolis and its hinterland, with particular emphasis on cult sites in Roman Thrace. Using a stable isotope (δ13C and δ18O) and trace element analysis via ICP-MS, integrated with petrography, the provenance of marble used in these artefacts was determined. The results identify Asenovgrad marble—prevalent in the Rhodope Mountains—as the primary material, with a minority of examples sourced from Prokonnesos and the southeastern Rhodopes. The predominantly fine-to-medium-grained nature of Asenovgrad marble may have facilitated the execution of small, detailed iconography. The consistent use of this local stone across votive types and deities implies centralised production, likely by one or more local workshops. These findings contribute to our understanding of marble selection, workshop localisation, and regional connectivity during the Roman period. They also highlight the importance of combining geochemical and petrographic methods for reliable marble provenance. This research demonstrates that even modest sanctuaries can yield critical data on raw material distribution and artisanal practices within broader economic and cultural frameworks.

1. Introduction

Throughout history, sanctuaries have served as focal points for artistic expression, community devotion, and material exchange [1]. The study of artefacts from these sacred sites offers invaluable insights into regional economic structures, workshop traditions, and the movement of both materials and artisans. In Roman Thrace, the sanctuaries played a vital role in maintaining local religious customs while simultaneously participating in broader artistic and economic networks [2]. Among the diverse objects associated with these sites, marble votive reliefs stand out due to their intricate craftsmanship and the potential they hold for provenance analysis. The case study of the sanctuary of the Nymphs from the village of Ognyanovo near Plovdiv (ancient Philippopolis) is an important case study for material provenance research, with the numerous marble votive reliefs found on-site offering a rich dataset for material provenance studies (Figure 1a). This research seeks to answer several key questions: Where did the marble originate from? What do the results reveal about workshop practices and the organisation of sculptural production? How were these votive reliefs connected to Philippopolis, and what can they tell us about potential regional stone trade networks? Their provenance results contribute significantly to understanding local workshop traditions, trade networks, and the sourcing of raw materials in the region.
Excavations of the sanctuary began in 1896 by V. Dobruski, who uncovered part of a building built of broken stones and mortar, from when the most significant number of votive reliefs were found [3]. Later, in 1987 and 1988, complete archaeological investigations of the complex were carried out. This complex probably included a complex of buildings, and cultic activity was carried out outside the sanctuary. The dedicatory inscriptions from the sanctuary honour the three nymphs bearing the local epithet Βoυρδαπηναι, which is derived from the name of the nearby settlement of Burdapa and characterises the nymphs as its patronesses [4].
Iconographically, the monuments can be divided into two major groups: the nymphs are depicted either dressed in long chitons and facing each other, or as the Three Harites, shown nude and dancing. The style and treatment of the forms and volumes suggest that they were made in a workshop near the shrine. The iconographic schemes preserved in the sanctuary at Ognyanovo lack notable local distinctions. The diversity of the type is noticeable in some details that enrich it, but all of them emphasise the aspect of the goddesses as protectors of water and springs. The sanctuary was probably functioning by the mid-3rd century AD, as evidenced by the numismatic evidence [5].
In recent years, several archaeometric studies have significantly advanced our understanding of marble use in Roman Thrace, with particular emphasis on urban centres such as Philippopolis, Montana, etc. These works have identified major quarry sources—most notably Asenovgrad, SE Rhodope, and Prokonnesos—and have begun to explore the organisation of local workshops and the circulation of materials in both public and funerary contexts [6,7,8]. Methodologically, the field has progressed through the combined application of stable isotope analysis (δ13C and δ18O), petrography, and trace element analysis via ICP-MS, supported by large-scale comparative databases [7,9,10,11,12]. Building upon this foundation, the present study contributes new data by focusing on the material composition of votive reliefs from rural sanctuaries in the territory of Philippopolis—an object group that has received limited dedicated attention. In addition to reinforcing the role of Asenovgrad marble in regional production, this study integrates petrographic evidence with statistically supported geochemical provenance to shed light on workshop dynamics, material standardisation, and the interconnectedness of rural cult practices with urban artisanal centres.

1.1. Geological Setting

The marbles of Asenovgrad exhibit petrographic and textural characteristics typical of high-grade metamorphic rocks within the Rhodope Mountains. According to Kovachev [13], these marbles form part of the Dobrostan Marble Formation, a stratigraphic unit within the Asenovgrad Group, composed of Precambrian metamorphic sequences. The formation comprises layered carbonate rocks interbedded with schists, kyanite–staurolite gneiss–schist, and amphibolites—clear indicators of intense metamorphism. The Dobrostan Marble Formation is structurally associated with the North Rhodope Anticline, where it overlies the Lukovitsa Gneiss–Schist Formation and is overlain by the Belashtitsa Carbonate–Silicate Formation, forming a distinctive and traceable anticline fold structure [14]. Marbles are particularly abundant in the upper sections of the Lukovitsa Formation and show lateral continuity with substantial thicknesses, often exceeding several tens of metres.
This formation is subdivided into three levels:
  • The lower level contains grey-to-grey-white-banded medium-grained marbles interbedded with gneisses and schists.
  • The middle level comprises white-to-grey fine-to-medium-grained massive marbles containing graphite, mica, and quartz inclusions.
  • The upper level consists of dolomitic marbles that gradually transition into purer calcitic marbles. The formation as a whole reaches thicknesses of up to 1600 m [13].
Outcrops near Sitovo village expose the southwestern extension of these marble layers, particularly the Lyaskovo Member of the Lukovitsa Formation, while other sections (e.g., near Oreshets and Mostovo) feature distinctive brecciated marbles with a red matrix, likely formed by karst processes rather than tectonic activity. Modern quarrying has significantly altered the landscape, obscuring traces of Roman exploitation. Nonetheless, small ancient pits and dumps still testify to substantial historical marble extraction. Present-day Philippopolis (Plovdiv) is geologically framed by seven hills of Upper Cretaceous syenite, yet its marble artefacts—especially votive reliefs—were largely sourced from the nearby Asenovgrad marbles [13].

1.2. Remarks on the Sanctuaries

Material studies of artefacts from the sanctuaries of Roman Thrace have not yet been the subject of a comprehensive specialised study. Chronologically, the Roman sanctuaries encompass the period from the annexation of Thrace in AD 46 to the late 4th century AD, when many sanctuaries ceased functioning as a result of the Christianisation policies of the Empire. Archaeological and numismatic evidence suggests that some sanctuaries show continuity from the Hellenistic period. Most sanctuaries in Thrace are situated in flatlands, often at the edges of lowlands or foothills, and show a clear tendency to cluster along key communication routes—particularly the arterial roads leading to and from Philippopolis. While the mapping of known sites outlines several notable gaps—such as the northern territory of Philippopolis, parts of the Strandzha and Sakar mountains, and regions between key urban hubs—these “voids” are now understood to result largely from uneven research coverage and limitations in recognising cult sites without systematic excavation [15]. This is evident in areas such as Malko Tarnovo, Gledachevo, and Krepost, where conventional votive assemblages are absent or sparse. Topographically, sanctuaries occupy a wide range of settings—from open plains to broken terrain, and from areas on top of prehistoric mounds to elevated ridges. Many are placed just below mountain summits, a position believed to provide ritual visibility over the territory and symbolically extend divine protection across the associated communities. Such variation reflects the complex interplay between religious ideology, landscape visibility, and territorial identity in the region’s sacred geography. This broader sacred landscape is critical for understanding the provenance data in this study, as each sanctuary may have followed distinct procurement pathways influenced by geography, accessibility, and cultic significance. The revised manuscript now includes these contextual insights to better situate the analysed votive reliefs within the dynamic religious and economic landscape of Roman Thrace [15].

2. The Case Studies and the Sampled Artefacts

For this study, a carefully curated selection of 18 votive plates exhibiting the most representative features was chosen for detailed analysis (Table 1 and Table 2). Additionally, a broader collection of miscellaneous votive plates from Philippopolis was examined to provide a comparative perspective. The selection of artefacts from the sanctuary of the Nymphs at Ognyanovo was conducted in collaboration with archaeologists in charge of the collection. Particular emphasis was placed on iconographic diversity, ensuring that stylistic variations could be assessed against potential material differences. This approach allowed us to investigate whether different iconographic treatments might correspond to distinct marble sources. The same rationale guided the selection of the miscellaneous votives, following initial agreements with local authorities that future research could extend the dataset to encompass additional votive reliefs if needed.
It should be noted that provenance marble studies of the statues and reliefs from the sanctuaries of Roman Thrace have not yet been conducted [15]. In this context, the selection strategy also considered practical aspects such as the physical condition of the reliefs; only those sufficiently intact and not overly weathered were eligible for sampling. The chronological frame was defined by the parameters of the FWF-funded project, focusing on the 1st–3rd century AD. While the material was initially introduced with an iconographic overview, all artefacts were documented according to their archaeological context, and the discussion below now foregrounds provenance by sanctuary or findspot. This adjustment better reflects the procurement dynamics specific to each cult site or urban context. A more detailed study on regional workshops is currently in preparation, and future research may pursue broader permissions for expanded sampling and comparative analysis.
The remaining 24 votive reliefs depict various deities such as the Thracian Horseman, Mithras, Zeus and Hera, Hermes (and Athena), Dionysus, Asclepius, Hygeia (the Asclepius votives from Philippopolis are left unfinished, and it is an indication that the preparation was performed in situ [16]), and Hekate in multiple sizes, and are linked to various shrines of the city and the vicinity. Several cult sites are associated with Zeus and Hera in the territory of Philippopolis; they were worshipped both independently and with a common cult. The popularity of the independent cult of Hera over that of Zeus or their common cult finds confirmation among the numerous single finds of votive tablets that are dedicated to the goddess. The mapping of their shrines shows their concentration in the lands west of Philippopolis and coincides with the spread of the cult of Asclepius in Thrace [15] (p. 295). The recorded shrines are mainly in the urban area of Philippopolis. Zeus and Hera are depicted in the identical iconographic scheme popular in Thracian lands during the Roman era. Zeus is depicted standing upright wearing a himation with a patera in his lowered right hand and a staff in his left. An altar and an eagle were often added. The image of Hera depicts the goddess standing upright, wearing a long chiton and a himation with a patera in the right hand and a staff in the left [15] (p. 122).
Major urban centres are the main centres of the Hermes cult. It is possible to take this fact as evidence of a foreign religious influence that penetrated Thracian lands through Roman and Greek urban cultures. With the help of quantitative and geographical-territorial analysis, the Hermes–Mercury cult’s main centres in Thrace’s interior can be distinguished. These are the administrative territories of Nicopolis ad Istrum, Pautalia, especially Philippopolis and Augusta Traiana, together with the middle reaches of the Hebrus and Tonzus rivers. The image of Hermes on the votive reliefs has a standard iconographic scheme, depicted as erect, frontally facing, half-naked, with a chlamys descending over his back. On the one hand, he holds a caduceus. He is often depicted with Athena, and in a monument from Philippopolis, Fortuna is represented on his left side [17].
So far, no archaeological sanctuaries of Dionysus have been excavated in Thrace, but several sites can be located based on the votive reliefs found. On the territory of Philippopolis, the cult of the deity is attested at Bozukova Mandra and Brani Pole [18] (pp. 424–425, No. 2, Figure 2); [15] (p. 129).
The most numerous sanctuaries recorded in the province of Thrace belong to the so-called Thracian Horseman. The mapping shows several regions with a concentration of cult sites. These are the areas of Nicopolis ad Istrum and Marcianopolis north of the Haemus Mountains. Here, the clustering of sanctuaries, mainly near the two urban centres, is striking. The network of cult sites is densest within the territories of Augusta Traiana and Philippopolis, where they are relatively evenly distributed. The popularity of the cult in the territory of Serdica and Pautalia is considerably weak. The topic of the Thracian Horsemen has been extensively studied by Boteva D. (e.g., [19,20]). Valchev [15] lists a total of fifteen shrines to the Thracian Horseman that have been found in the urban territory of Philippopolis (an indication of a marble statue workshop is located on Gladstone Street, Plovdiv, with semi-finished objects of the Thracian Horseman [21] (p. 297)). In the monuments found south of Haemus, the type of the horseman calmly moving to the right, very often accompanied by a female figure depicted standing in front of the horse, prevails, while in the north—in the territory of Marcianopolis—the image of the hunting horseman is preferred [15] (pp. 112–113).
In Thrace, the cult of Mithras spread mainly in the urban areas of Serdica, Pautalia, and Philippopolis. The god gained particular popularity in the cities of Serdica and Pautalia, where the existence of a Mithraeum was assumed. The main route of his penetration into Thracian lands was with mercenaries in the Roman army [22].
Table 2. The sampled miscellaneous artefacts from Philippopolis and its region.
Table 2. The sampled miscellaneous artefacts from Philippopolis and its region.
Project NrLab NrM. Nr *ObjectMaterialStone DescriptionDatingLocationFind SpotDetails on Finding-SpotProvenance ResultsPhotoBibliography
FWM004792431537Votive plateMarbleWhite colour; ultrafine-grained.2nd century ADNAIM MuseumPhilippopolisn/aAsenovgradMinerals 15 00670 i019IGBulg III,1 965
Description: Relief image of Zeus and a Greek inscription.
FWM004992441643Votive plateMarbleWhite colour; coarse-grained.2nd–3rd century ADNAIM MuseumAsenovgrad regionn/aAsenovgradMinerals 15 00670 i020[23] (Figure 444, p. LXXV)
Description: Relief image of the Thracian Horseman.
FWM010292672906Votive plateMarbleWhite colour; fine-grained with larger crystals.2nd–3rd century ADNAIM MuseumPanagyurishte

Philippopolis
n/aAsenovgradMinerals 15 00670 i021IGBulg III,1 1055
Description: Relief images of Dionysos and Heracles in a chariot. Greek inscription engraved on the frame.
FWM010492685993Votive plateMarbleWhite colour; fine-grained with larger crystals; heteroblastic; zero translucency.2nd–3rd century ADNAIM MuseumPhilippopolis, Philippopolis regionn/aAsenovgradMinerals 15 00670 i022[24]
Description: Relief images of Athena and Hermes.
FWM010592693337Votive plateMarbleWhite colour; coarse-grained.First half of the 3rd century ADNAIM MuseumTatarevo,

Philippopolis region
n/aAsenovgradMinerals 15 00670 i023IGBulg II 768
Description: Relief image of Zeus holding a horn of plenty and a patera over an altar.
FWM010892643310Votive plateMarbleWhite colour; medium-to-coarse-grained.2nd–3rd century ADNAIM MuseumPhilippopolis, Philippopolis regionn/aAsenovgradMinerals 15 00670 i024Note: Under the current investigation with the project FWF P 33.042
Description: Representing a scene of a sacrifice.
FWM011292651531Votive plateMarbleWhite to beige colour; coarse-grained.2nd–3rd century ADNAIM MuseumPhilippopolis, Philippopolis regionn/aAsenovgradMinerals 15 00670 i025[25]
Description: Relief images of Dionysus and Heracles on a chariot. Scene of wine production in the lower register.
FWM004192411642Relief of HerculesMarbleWhite-to-greyish colour; medium-to-coarse-grained.2nd–3rd century ADNAIM DepotAsenovgrad regionn/aSE Rhodopen/an/a
Description: A votive plate decorated with a relief representing a scene of sacrifice. Heracles (on the left) and another male figure (on the right, smaller in size) are shown standing at the sides of an altar, making a libation upon it. A tree is visible in the background.
FWM046098152102Votive plate of MithrasMarbleWhite colour; medium-grained.2nd century ADRAMP **Kurtovo Konaren/aAsenovgradMinerals 15 00670 i026[26]
Description: The relief contains a complex iconography of divine figures and symbols. The background in the centrepiece is cut, i.e., an ajour style is used. Traces of red paint are preserved on the surface of the relief. The central image represents Mithras killing a sacred bull, an act called the tauroctony. Mithras, wearing a Phrygian cap, is kneeling on the exhausted bull, holding it by the nostrils with his left hand and stabbing it with his right. A dog and a snake reach up towards the blood. At the top left is Sol, and at the top right is Luna. The central tauroctony is framed by a series of subsidiary scenes illustrating events in the Mithras narrative.
FWM04619816II-730Votive plate of the Thracian horsemanMarbleWhite/greyish colour; medium-to-coarse-grained3rd century ADRAMPPlovdivn/aAsenovgradMinerals 15 00670 i027[26]
Description: The plate has a trapezoidal shape, rounded at the top. A frame is formed at the bottom. Traces of red paint are preserved on the surface of the relief. The rider is turned to the right, and a chlamys flutters behind him. A hunting scene occurs under the images of the horse and the horseman. The Three Nymphs and a temple facade are depicted in the upper right corner of the plaque.
FWM046298172103Votive plateMarbleWhite/greyish colour; medium-grained.2nd–3rd century ADRAMPPlovdivn/aUnknown ***Minerals 15 00670 i028IGBulg III,1 959
Description: Depicts a three-headed Thracian horseman. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part. An inscription is incised in the broader part of the frame, and a small round hole is drilled through it. The three-headed rider is turned to the right, and a chlamys flutters behind him. He holds a double axe in his right hand. A hunting scene occurs under the images of the horse and the horseman. At the right end of the relief, a pedestal and two standing figures above it are depicted.
FWM04729827II-1026Fragment of a votive reliefMarbleWhite colour; fine-grained with larger crystals2nd century ADDepot of the RAMPPlovdivn/aAsenovgradn/an/a
Description: The centrepiece is set between two columns, and the shafts are broken off. Two Corinthianising capitals support the entablature. Herakles, on the right, is represented frontally, with his head turned to the right towards the Amazon. The skin of the Nemean lion is slung over his right arm. The Amazon is represented with her head tilted back. Her chiton is unfastened at one shoulder. Her left arm is folded up, holding a round object. There are traces of paint.
FWM047498292219Fragment of a votive plateMarbleWhite colour; medium-grained2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i029
Minerals 15 00670 i030
[27] (p. 46, Figure 36)
Description: An inscription is incised on the frame at the lower end of the plaque. A very small part of the relief is preserved, including a foot.
FWM047698313952Votive plateMarbleWhite colour; ultra-fine-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i031[28]
Description: Depicts a Thracian horseman. The plaque is surrounded by a frame that is wider at the lower part. An inscription is incised on the wider part of the frame. The upper part of the plate, together with most of the rider and the horse’s head, is broken off. The rider is turned to the right. The horse’s front left leg is folded and lifted into the air.
FWM047898334040Votive plateMarbleWhite colour; medium-grained.End of the 2nd–3rd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i032IGBulg III,1 1460
Description: Depicts a Thracian horseman. The plate has a trapezoidal shape, probably rounded at the top, where it is largely broken off. The plaque is surrounded by a frame that is wider at the lower part. An inscription is incised on the wider part of the frame. The rider is turned to the right, and a chlamys flutters behind him. He holds a spear in his right hand. The horse’s front legs are raised in the air. Below them, there is a shallow image of a dog. A pedestal or an altar is partially depicted in the lower right corner of the relief.
FWM04829837II-1579Votive plateMarbleWhite colour; ultra-fine-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i033[29] (p. 134)
Description: Depicts Asclepius, Hygieia, and Telesphorus. The plate has a trapezoidal shape, rounded at the top. It is surrounded by a frame that is wider at the lower part. On the right is Asclepius with a serpent, on the left is Hygia, and in the space above and between them is Telesphorus. The relief of the images is low; the figures are schematic, and no details are presented.
FWM04839838II-1545Votive plateMarbleWhite colour; ultra-fine-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i034[29] (p. 133)
Description: With two divine figures. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part. On the right is a goddess (Hera/Tyche) depicted with a high cylindrical crown (kalathos?) and cornucopia on the left hand and steering oar in the right, and on the left is a god (Hermes) depicted with a purse in his right hand and a stylised caduceus (?) in his left hand. The relief of the images is shallow, and no details are presented.
FWM04849839II-1546Votive plateMarbleWhite colour; fine-to-medium-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i035[29] (p. 133)
Description: Depicts Hermes and Hera. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part. On the right is Hermes, depicted with a purse in his right hand, and on the left is Hera, depicted holding a patera in her right hand and a sceptre in her left hand. The two figures are presented frontally. The relief of the images is low; the figures are schematic, and no details are presented.
FWM04859840II-275Votive plateMarbleWhite colour; fine-grained with larger crystals.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i036[29] (p. 135)
Description: Depicts Hera. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part. An inscription is incised on the wider part of the frame. The relief depicts Hera holding a patera in her right hand and a sceptre in her left hand. The figure is presented frontally. Beneath the patera, in the lower left corner of the relief, there is an altar. The relief of the image is low; the figure is schematic, and no details are presented.
FWM04869841II-36Votive plateMarbleWhite colour; coarse-grained2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i037[29] (p. 135)
Description: Depicts Hekate. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part.
FWM04879842II-1758Votive plateMarbleWhite colour; fine-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradMinerals 15 00670 i038[29] (p. 135)
Description: Depicts Hera and is split into two. The plate has a trapezoidal shape, rounded at the top. The plaque is surrounded by a frame that is wider at the lower part. The relief depicts Hera holding a patera in her right hand and a sceptre in her left hand. The figure is presented frontally. Beneath the patera, in the lower left corner of the relief, there is an altar. The relief of the image is low; the figure is schematic, and no details are presented.
FWM05049859n/aFragment of reliefMarbleWhite colour; medium-grained.2nd–3rd century ADDepot of the RAMPPlovdivn/aAsenovgradn/an/a
Description: Only the right leg of a man is preserved.
FWM05059860n/aVotive reliefMarbleWhite colour; fine-grained.2nd century ADDepot of the RAMPPlovdivn/aAsenovgradn/an/a
Description: Possibly a small statue of Cybele. The sculpture depicts the seated goddess, Cybele, the Mother Goddess (Magna Mater), with two flanking lions. Cybele’s head and part of her arms are broken off. The lower part of her left arm is placed on an armrest. The goddess is wrapped in heavy drapery. Each lion sits by her side. The head of the left one is broken off. The goddess is presented on a much larger scale than the wild animals.
FWM047798322241Votive reliefMarbleWhite colour; medium-grained.2nd–3rd century ADDepot of the RAMPPlovdivn/aProkonnesosMinerals 15 00670 i039IGBulg III,1 1133
Description: Depicts Asclepius and Telesphorus. The small-sized sculpture represents the figures of Asclepius in the middle, Telesphorus on the right, and the entwined snake on a baton on the left. The head, torso and arms of Asclepius, as well as the head of Telesphorus and the upper part of the serpent, are broken off. The two gods are depicted with bare feet. The figures are placed on a plain rectangular base on which an inscription is incised.
* Museum number. ** Regional Archaeological Museum Plovdiv (RAMP). *** While the geochemical profile of the sample closely matches that of Asenovgrad marble, minor discrepancies suggest it may originate from an as-yet-unidentified source—possibly a local quarry or outcrop in the vicinity.

3. Methodology

Studies of ancient stone materials and their provenance offer critical insights into historical trade networks, construction practices, and workshop organisation. Examples from diverse material categories—such as volcanic and pyroclastic rocks, pozzolanic binders, limestones, and marbles—have demonstrated the value of archaeometric approaches in tracing sources and understanding technological choices in antiquity. Recent research has highlighted the efficacy of combining trace element analysis with petrographic and isotopic techniques to determine provenance, particularly within Roman archaeological contexts [30,31,32,33,34]. These methodologies parallel our own multi-proxy approach, which integrates ICP-MS trace element analysis, stable isotope ratios, and petrography to establish the origin of marble used in votive reliefs with high confidence. While the present study centres on marble, comparative work on other lithologies affirms the broader applicability and reliability of trace element analysis in provenance investigations. Including such case studies not only reinforces the methodological soundness of our approach but also positions this research within the wider framework of established archaeometric practice.
In recent years, systematic research efforts have significantly expanded the known corpus of ancient white and whitish marble quarries. These discoveries have not only enriched our understanding of marble use and distribution across different historical periods, but have also driven advancements in analytical methodologies, database development, and the refinement of provenance discrimination techniques. The continuous identification of new quarry sources necessitates a more comprehensive approach to tracing marble’s historical trade and utilisation patterns. Various research teams worldwide have engaged in marble provenance studies, employing diverse analytical methodologies and focusing on different quarry regions [35,36,37,38,39,40,41,42]. In this broader context, our team has developed an extensive comparative collection of geological samples (see below), all analysed using a consistent multi-method approach. This allows for direct comparability with archaeological artefacts examined using the same protocol.
For this study, archaeological samples were systematically compared against an extensive reference database containing 5700 geological samples from ancient marble quarries and outcrops. This dataset encompasses well-documented sources such as Aphrodisias, Carrara, Prokonnesos, Dokimeion, Penteli, Paros, Thasos, Heraclea/Miletos, and Ephesos, as well as various sites in Roman Macedonia and Thrace, among others (https://oeaimetrix.oeaw.ac.at/, accessed 6 June 2025). Additionally, quarries and outcrops from surrounding regions were incorporated to ensure a thorough comparative analysis. The following diagrams illustrate the final selection of quarries that exhibit the closest geochemical similarities to the studied archaeological specimens. Given the complexity of marble provenance studies—marked by substantial geochemical variability and the vast number of potential quarry sources—reliable provenance determination requires an integrated approach that combines multiple analytical methods. Once the measurements are completed, statistical analysis is applied, allowing for direct comparisons with established databases to accurately determine the marble’s origin.
The methodologies and the geological data employed in this study follow those detailed extensively in [7,9,10,11,12,43]. Accordingly, only a summarised version of the procedures and relevant updates is provided in the section below (The study generated numerical data for the Philippopolis samples using both stable isotope analysis and ICP-MS analysis methods. Stable isotope and ICP-MS analysis were employed to acquire comprehensive insights into the composition and characteristics of the samples. The utilisation of these advanced analytical techniques facilitated a detailed examination of the isotopic signal and elemental composition of the samples, thereby enhancing the precision and depth of the data obtained. The data are presented with the lab and sample number including the stable isotopic values of Carbon and Oxygen and the chemical values from the ICP-MS analysis (magnesium, manganese, iron, strontium, chromium, vanadium, yttrium, cadmium, boron, lanthanum, cerium, praseodymium, dysprosium, holmium, ytterbium, lead, and uranium, data in ppm)).

3.1. Microscopy

Petrographic features are of limited use due to the textural variability of the marble samples, even within a single quarry. However, significant petrographic differences can be seen between the marble groups examined in research, allowing for a priori categorisation into the groups [44,45,46].
The analyses below were conducted using a petrographic microscope, with samples prepared as thin sections and inspected under a Kern OPO 185 microscope.

3.2. Isotope Analysis

The examination of stable oxygen (O) and carbon (C) isotopes is a widely accepted method in marble provenance studies, with numerous laboratories offering this analytical service. While the technique is considered to be highly reliable, careful sampling is required when dealing with weathered or contaminated surfaces to ensure accurate results. Proper precautions must be taken to avoid alterations that could impact the isotopic signature and lead to misinterpretation [47,48,49,50].
The current samples were analysed at the Bayerische Staatssammlung für Paläontologie, Munich, Germany. Carbon dioxide was extracted from carbonates by reaction with phosphoric acid at 72 °C using an automated online device (Finnigan Gasbench II). Oxygen and carbon isotope ratios were measured with a Finnigan Delta-plus mass spectrometer operated in continuous He flow mode. A house standard was used as the running standard.

3.3. Trace Element Analysis

In this study, trace element concentrations were determined using ICP-MS (Inductively Coupled Plasma Mass Spectrometry) following the selective dissolution of the carbonate phase with hot HNO₃. This method specifically targets the carbonate component without dissolving silicates, making it crucial that the obtained data are not directly compared or combined with the results from bulk analytical techniques, such as HF dissolution or XRF analysis. Given that many key trace elements in marble—such as Mn, Mg, Fe, Sr, Y, V, Cd, Ba, La, Ce, Yb, and U—are present only in sub-ppm concentrations, meticulous sample preparation and analytical precision are essential [35,51,52].
For the presented study, the analysis was conducted using an Agilent 8800 ICP Triple Quad (ICP-QQQ) mass spectrometer calibrated against a Merck VI standard. Limestone JLs-1 served as the internal standard during the process. For this investigation, the samples were analysed at the Chemistry Department of TU Wien.

4. Provenance Results

4.1. Petrographic Data

Due to the limited sample size, only three thin sections were prepared; nevertheless, distinct characteristics of the local marble are clearly observable. The three votive reliefs, FWM0047 (Figure 2a,b), FWM0104 (Figure 3a,b), and FWM0353 (Figure 4a,b), show similar characteristics with fine-to-medium grain sizes. All three samples have an ultra-fine-mylonitic matrix (blastomylonitic groundmass) with larger pre-deformed relics (with corroded surroundings). Undulose extinction can be observed on the grains, a characteristic often seen in the Rhodope marbles. The fabric features of the first two samples have a mosaic feature with straight-to-curvy grain boundaries, while the third has a mortar fabric feature with sutured boundaries. To support this comparison, a thin section from a quarry sample (FWM0846) attributed to Asenovgrad marble was also examined (Figure 5a,b), showing analogous textural traits and confirming the petrographic consistency between the quarry material and the artefacts.
The marbles of Asenovgrad share many textural features with other important quarry sites of the Rhodope Mountains. According to Kovachev [13], these marble locations belong to a series of Precambrian high metamorphosed formations, where layered carbonate series are intercalated in layers of schist, kyanite–staurolite gneiss–schist, and amphibolites, proving the high-grade metamorphic grade of these rocks (Twin lamellae are observed in some of the votive samples (e.g., FWM0353), where they appear slightly curved, although not as distinctly deformed as in the quarry reference sample (FWM0846; Figure 5). The limited visibility of this feature in certain artefacts may be due to surface alteration, section orientation, or sample preservation. Nonetheless, the general textural features remain consistent with Asenovgrad marble.
These analyses were conducted using a petrographic microscope, with samples prepared as thin sections and inspected under a Kern OPO 185 microscope. The comparative analysis relied on the thin-section database established by Prof. Prochaska (ÖAI/ÖAW), which includes geological samples from both ancient and modern quarries and is available from the Austrian Archaeological Institute (approximately 350 samples). Examples of comparative thin sections can be found in [11], while further petrographic information specifically related to Asenovgrad marble artefacts is available in [23]. Granulometric features were recorded during macroscopic and petrographic analyses and are summarised in Table 1 and Table 2. Remarks on grain sizes: 0.5–1.2 mm—fine grain; 1.2–2.5 mm—medium grain; >2.5 mm—coarse grain [7].

4.2. Geochemical Data

Isotopic and multivariate diagrams unveil preliminary findings, elucidating diverse local, regional, and supra-regional sources. The isotope diagram proves challenging to interpret definitively due to the intertwining ellipses from multiple sources, casting uncertainty on any conclusive findings (Figure 6). The multivariate diagram provides more apparent discrimination by using the variables Mn, Mg, Fe, Sr, Cr, V, Ba, Y, Cd, La, Ce, Yb, U, δ18O‰, and δ13C‰ (Figure 7). However, there is also an intersection among specific sources.
Employing SPSS (version 30.0.0) with consistent variables furnishes numerical probabilities and associated outcomes to facilitate clear differentiation (Table 3). To handle the complexity inherent in the acquisition of multiple geochemical variables, a systematic and rigorous statistical framework was implemented. Integrating the results of stable isotope and trace element analyses necessitated a carefully calibrated workflow grounded in established statistical principles. The research team applied a multivariate discriminant analysis to assess and interpret the dataset, drawing on the capabilities of software platforms such as STATISTICA and SPSS—tools that have demonstrated reliability in similar archaeometric contexts (cf. [7]). These complementary software environments enabled both the graphical and quantitative differentiation of the data. In SPSS, discriminant functions were generated to statistically allocate each sample to a defined quarry group based on its elemental signature, producing probabilistic assignments that quantify classification confidence. In parallel, STATISTICA was used to construct a set of synthetic discriminant factors, combining multiple chemical variables into reduced dimensions that allow for the visual interpretation of sample clustering. These plots serve as diagnostic tools for evaluating the geochemical correspondence between archaeological artefacts and potential source quarries. The same set of variables was uniformly applied across all datasets to maintain methodological consistency (see Table S1). This approach permitted a direct comparison with an extensive reference database, thereby allowing for the confident determination of marble provenance for each artefact. The outcome is a robust classification model that minimises the scope for subjective interpretation and enhances reproducibility across datasets. By employing this protocol, this study offers a reliable means of distinguishing between different quarry sources. The strength of the model lies not only in its statistical rigour, but also in its capacity to synthesise complex data into accessible formats—both numerical and visual—ultimately allowing for a high degree of certainty in provenance attribution.
The artefacts are separated into two groups. The blue points are votive reliefs depicting different deities and originating from various find spots, while the second group (red points) comes from one location, the Sanctuary of the Three Nymphs. The majority of the results show the origin of local marble from Asenovgrad for the artefacts. Two of the objects are made of Prokonnesian marble (FWM0477; FWM0348), while one of the objects is made of SE Rhodope marble (FWM0041). The FWM0348 is the only artefact from the Sanctuary of the Three Nymphs made of Prokonnesian marble. The rest of the votive plates are exclusively made of Asenovgrad marble. The artefact FWM0462 remains unknown. However, the majority of its geochemical characteristics are similar to the Asenovgrad marble. This object may originate from a pit near the quarries that have not yet been discovered or analysed (Table S1).
An explanation of the table is provided below:
  • Distance: Distance of the sample under consideration from the centre of the ellipse. This centre is the average value of the quarry probability field.
  • Relative (posterior) probability: This probability is the degree of likelihood of a sample belonging to a given group (within the selected number of groups). Results below 60% indicate that the sample probably cannot be assigned with certainty and that a second choice has to be considered.
  • Absolute (typical) probability: This is the measure of the probability that a sample belongs to a given population. Samples in the centre of the probability ellipse have a high absolute probability. The threshold is 10%, corresponding to samples on the edge of the 90% probability ellipse. Low values indicate anomalous samples (outliers) or samples possibly not belonging to any group in the selection.
As noted, low absolute probability values (e.g., ≤10) indicate a greater statistical distance from the reference group and should be interpreted with caution. However, in such cases, the relative probability still enables a confident attribution to the most likely quarry, particularly when supported by consistency across multiple geochemical variables. In this study, most of the archaeological samples exhibit geochemical profiles closely aligned with the Asenovgrad reference group (see Table S1). Minor deviations may reflect the existence of currently undocumented or unsampled quarry pits in the broader Asenovgrad area, which merits further investigation. As described in the methodology (Section 3), the artefacts were compared against an extensive reference database of ancient marble sources. Only the final selection of quarry groups yielding statistically significant matches—namely, Asenovgrad, SE Rhodope, and Prokonnesos—is presented in Table 2 for clarity and focus.

5. Discussion

5.1. Materials

The analysis of votive reliefs and sanctuary artefacts from Philippopolis and its surrounding areas reveals a strong preference for locally sourced Asenovgrad marble. This pattern is evident across various artefact groups, including the votive reliefs dedicated to different deities and the sculptural representations from the Sanctuary of the Three Nymphs in Ognyanovo.
A. Reliefs from the sanctuary of the nymphs: The same pattern of material used can be seen in the artefact group of the Sanctuary of the Three Nymphs in Ognyanovo near Philippopolis. Despite the different stylistic approaches to the appearance of the Nymphs, all of the objects are made of Asenovgrad marble (Figure 8). There is only one exception: a poorly preserved artefact made of Prokonnesian marble. However, its stylistic characteristics and the material suggest an unfinished or badly worked object.
B. Votive reliefs from Philippopolis and its vicinity: A vast number of votive reliefs from Philippopolis and its surrounding region showed the almost exclusive use of Asenovgrad marble. The fine-grained consistency of the local marble possibly assisted in creating small-sized, detailed figures in each relief. Based on the same local source used for most objects with different deities, it can be assumed that they were produced in a local workshop/workshops.
There are only two exceptions in the analysis of twenty-four votive reliefs, with the use of Prokonnesos (FWM0348) and SE Rhodope (FWM0041) (Figure 9). A similar hypothesis of the inscribed objects, which can be developed from the current study, is that they used waste material from larger blocks to create new objects, such as small votive plates. The availability of such waste material can be considered to be high, especially in relation to the extensive architectural programmes carried out in Philippopolis, where large-scale imports and the shaping of blocks would likely generate offcuts. Of course, it cannot be disregarded that possibly pre-shaped plates were imported. However, the fact that, in a large group of artefacts, only two are made from imported marble suggests circumstantial or opportunistic marble use rather than a systematic supply from Prokonnesos or SE Rhodope. When compared to broader regional dynamics, the data from the Lozen sanctuary and other parts of Thrace support the notion that local marble sources like Asenovgrad were critical to intra-provincial trade and the operation of workshop economies. The infrequent appearance of imported marbles (e.g., Prokonnesos or SE Rhodope) in votive reliefs—often associated with stylistic anomalies or lower-quality workmanship—underscores the strength of local supply chains and their competitiveness in both quality and accessibility. In support of the reuse hypothesis, it is worth noting that similar patterns have been observed in other regional contexts, such as at the sanctuary of Apollo Γεικεσηνος at Lozen, where a number of reliefs made from non-local marbles appeared to be stylistically or materially inconsistent with workshop standards, suggesting the secondary use or local reworking of pre-existing materials [53].
The votives of Asclepius can be found across other cities, such as Nicopolis ad Istrum and Dionysopolis, as well as the sanctuaries of Batkun and Glava Panega [54]. Future analysis will show us if the material used for these reliefs is the same as in Philippopolis, made of Asenovgrad marble, or if this pattern was used with various materials. Similarly, further analysis of votive plates from other deities, such as Thracian Horseman, Mithras, etc., from across the province will enlighten us as to what the material preference for this type of object was.
In Ancient Thrace, marble quarries were typically located near settlements, facilitating both extraction and trade through well-established routes. This pattern can be observed at several key sites, including the Berkovitsa quarry, situated approximately 25 km from Montana, the Asenovgrad quarries about 20 km from Philippopolis, and the SE Rhodope marble sources, which supplied material to sites such as Villa Armira, located between 2 and 12 km away [7]. Similarly, major limestone quarries were strategically positioned near urban centres, such as Samovodene, around 10 km from Nicopolis ad Istrum [55], and Marciana, approximately 5 km from Marcianopolis [6]. The accessibility of both local and regional stone resources reinforced Philippopolis’s role in the Roman Empire’s trade and architectural networks. The combination of locally sourced and imported materials reflects a sophisticated economic system that supported extensive construction and artistic production in the region.
Asenovgrad, located in central Bulgaria, was one of the key marble extraction sites in Thrace. Four quarries, known as Ravdin 1, 2, 3, and 4, were analysed, all of which are situated in close proximity and yield white marble with medium-to-coarse grains interspersed with finer-grained layers. While signs of modern extraction are present at each site, no definitive traces of ancient quarrying activity have been identified. However, artefactual analysis from various locations across the Thracian province [6,8] confirms that these marble sources were actively exploited during the Roman period. Marble from Asenovgrad was extensively used in both sculptural and architectural contexts in and around Philippopolis during the Roman period. Notably, several sculptures sampled from the city—including portraits, imperial figures, and deities—were carved from this local material, reflecting both its technical suitability for detailed work and its economic accessibility [8]. A significant portion of the artefacts examined in the present study are composed of Asenovgrad marble, further underscoring its widespread use in ancient craftsmanship and construction.

5.2. The Chronological Distribution

a. 2nd century: The largest group of votive reliefs, made of Asenovgrad marble, dates to the 2nd century AD. Only two samples dated from the 2nd to 3rd century AD are made of sub-regional and supra-regional quarries, SE Rhodope (FWM0041) and Prokonnesos (FWM0477). As explained above, the material from these two objects may come from marble waste of larger blocks (Figure 10).
b. 3rd century: In the middle of the 3rd century AD, the votives related to the Sanctuary of the three nymphs in Ognyanovo are made of Asenovgrad marble.
The chronological distribution of the sampled artefacts, as outlined in this study, reveals a marked preference for Asenovgrad marble during the 2nd and 3rd century AD. This is a period of intensified urbanisation and cultic investment in Roman Thrace, particularly in the Philippopolis region [56]. The predominance of this local marble in both public and private cultic contexts suggests a deliberate and sustained exploitation of regional resources by local workshops, reinforcing the hypothesis of standardised production models within a highly integrated local economy. From a workshop production standpoint, the use of Asenovgrad marble across different deities and sanctuary sites—especially the near-universal application within the Sanctuary of the Three Nymphs—indicates the operation of established carving traditions and logistical networks. From a historical and economic perspective, this reflects a strategic integration of raw material procurement and artisanal output within a provincial framework. Local quarries, such as Asenovgrad, not only supplied material to nearby urban centres like Philippopolis but also shaped regional identity through recurring iconographic and material patterns. Their exploitation during the 2nd and 3rd centuries parallels the broader Roman imperial economic policy of developing internal resources to reduce dependency on costly imports, particularly in areas where high-quality local alternatives existed. Future comparative analyses, particularly across other urban and rural sanctuaries in Thrace, will further clarify the extent to which this local marble economy influenced regional workshop distribution, trade circulation, and material choices over time.

5.3. The Use of Asenovgrad Marble in Other Cultic Contexts and Workshops

Beyond the Sanctuary of the Three Nymphs, Asenovgrad marble played a central role in other religious contexts in the Thracian region. A notable example is the sanctuary of Apollo Γεικεσηνος, situated in the southeastern area of the province, where recent archaeometric investigations have confirmed that the majority of votive reliefs were likewise carved from local Asenovgrad marble [53]. Although this sanctuary was not discussed earlier in detail, its inclusion here serves to illustrate the broader pattern of reliance on local marble resources across various cultic sites in Roman Thrace. Several objects crafted from Asenovgrad marble exhibit stylistic coherence—especially those depicting the Thracian Horseman in battle scenes or with elaborate iconographic schemes—indicating production within a single workshop or closely affiliated ateliers. Furthermore, the identification of inscriptions and dedications by Roman soldiers, including those bearing ethnonyms traceable to Thracian villages in the Philippopolis territory, supports the hypothesis that the sanctuary’s reach extended well beyond its immediate locale, drawing upon urban resources both material and artisanal. This underscores the strategic role of Philippopolis and its marble quarries in supplying and shaping regional cultic landscapes through material and artistic production [53].
Located near the ancient city of Philippopolis, Asenovgrad’s marble was extensively exploited during the Roman period, serving as the primary source for a significant number of artefacts found at various sites [6], including those previously examined. The high frequency and quality of votive monuments made from this marble not only suggest its aesthetic and technical suitability but also point towards the involvement of urban workshops in Philippopolis, which likely maintained regular supply and production chains for sanctuaries in the wider region.
A small set of votive reliefs from Philippopolis, found in a semi-finished state, reveals the existence of workshops for these types of objects in the city. Cherneva [29] presents a catalogue of these items in dedication to the deities of Hermes and Athena, Hermes and Fortuna/Tyche, Asclepius and Telesphorus, Hera, one relief of Hekate, etc. All of the mentioned catalogued objects were sampled and analysed, with the provenance results all being of Asenovgrad marble (FWM0482; FWM0483; FWM0484; FWM0485; FWM0486; FWM0487) (see above). Previous studies suggest a possible workshop located near the Eastern Gate of Philippopolis, where various types of objects, including the aforementioned votive reliefs and a grave stele, were carved [29,56]. Unfinished marble objects have been found in the area of Asenovgrad [57], which is a strong indication that production and possible workshops were established there [56].
These cumulative lines of evidence underscore the pivotal role of Asenovgrad marble not only as a preferred material but also as a cornerstone of a regionally interconnected system of cultic production, trade, and artistic expression in Roman Thrace.

5.4. Marble Transport and Regional Connectivity

The Ognyanovo sanctuary is located near the junction of two strategically important trans-provincial roads, providing excellent long-distance connections. Coming from the south, a Roman road connected the regional centre Nicopolis ad Nestum with the Hebrus Valley, Philippopolis, and the upper Thracian plain (Figure 1b). After passing through the mountains and past the village of Kapitan Dimitrievo, near Sinitovo, the westernmost Roman road through the Rhodope Mountains met the main trans-provincial road [58] (p. 73). Its ancient name is lost, but today it is known under the names via diagonalis or via militaris. Coming from Singidunum, it led toward via Naissus, Serdica, Philippopolis, and Hadrianopolis to the provincial capital in Perinthus. The harbour of Perinthus was home to a Roman fleet, the Classis Perinthia, and, furthermore, was a focal point for transhipments to Asia [59,60]. Marble from Prokonnesos could be transported along the via diagonalis or shipped upstream via the Hebrus river. On a regional level, the via diagonalis passing through the territory of Philippopolis provided the main thoroughfare and excellent conditions for the marble trade.
The marble quarries of Asenovgrad were connected to Philippopolis via a road that led through the Rhodope Mountains towards the Thracian Sea. Concerning the marble transport, oxen were the preferred draught animal for large and heavy loads. The problems of transport logistics were not primarily determined by the available tractive force, but rather by the stability of the transport medium and the nature of the terrain. The capacity of the loaded axle is the decisive factor [61] (pp. 591–592), [62] (p. 6). Given that Ognyanovo is located approximately 50 km west of Asenovgrad by road, it is estimated that an ox-drawn wagon travelling at a speed of 15–20 km per day would have required approximately three days to complete the journey [63] (p. 150) (Table 1).

6. Conclusions

This archaeometric study of the votive reliefs from the sanctuary of the Nymphs at Ognyanovo and from various sites in the territory of Philippopolis provides a robust framework for understanding the material and cultural dynamics of Roman Thrace. The overwhelming predominance of Asenovgrad marble across the analysed artefacts—especially those from the Sanctuary of the Three Nymphs—points to a deliberate and sustained use of local resources in the creation of these devotional objects. Only a very small number of exceptions, notably one relief made from Prokonnesian marble and another from SE Rhodope, highlight this pattern of local preference.
A significant outcome of this research is the evident iconographic consistency across the reliefs, despite some variations in stylistic detail. The recurring depiction of the Three Nymphs—either in a static pose or in the dynamic motif of the dancing Harites—along with standardised representations of major deities like Zeus, Hera, Hermes, and the Thracian Horseman, suggests a shared visual vocabulary and symbolic repertoire across the region. These similarities not only reflect a coherent religious narrative but also imply the existence of established artistic templates possibly circulated within a local network of artisans.
When combined with the strong geochemical correlation of the marble to the Asenovgrad quarries, the stylistic and iconographic parallels bolster the argument for the operation of one or more local workshops. These workshops likely had direct access to nearby quarries and served both urban and rural sanctuaries. Their production of votive reliefs was not only standardised in terms of material and format, but also deeply embedded in the socio-religious fabric of the region. Within this range, the 3rd century appears to mark a peak in production, particularly at the Sanctuary of the Three Nymphs. Although production may have begun earlier, the available material confirms that the local workshops were actively engaged during this time using Asenovgrad marble. From a historical-economic perspective, this concentration of activity reflects a period of intensified regional investment in cult infrastructure and artisanal output, supported by the accessibility of local resources and the integration of rural sanctuaries into the urban religious landscape.
This study underscores the intricate connection between material choice, artistic tradition, and workshop organisation in Roman Thrace. By anchoring votive production to local marble and recurring iconographic schemes, it becomes possible to more accurately trace the movement of artistic knowledge and the localisation of cult practices. Future comparative analyses across the wider province will be essential in confirming whether this pattern was uniquely characteristic of Philippopolis or is more broadly indicative of regional sculptural practices in Roman Thrace.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/min15070670/s1, Table S1: The complete geochemical dataset for all analysed artefacts, including trace element concentrations and stable isotope values used to determine marble provenance. Table S2: List of Roman-period sanctuaries in the territory of Philippopolis. The table includes geographical coordinates, associated deities, and references to mapped locations illustrated in Figure 1b.

Author Contributions

Conceptualization, V.A.; methodology, W.P. and V.A.; investigation, W.P., V.A., P.A., K.P. and B.F.; data curation, V.A.; writing—original draft preparation, V.A.; writing—review and editing, V.A., W.P., P.A., K.P. and B.F.; supervision, W.P.; Visualisation, B.F. All authors have read and agreed to the published version of the manuscript.

Funding

The project was sponsored by the Austrian Science Foundation, FWF (Österreichischer Wissenschaftsfonds), project P 33042 “Fingerprinting White Marbles. Quarries and Cities of Roman Thrace, 1st–3rd century AD”, awarded to Sabine Ladstätter between 2020 and 2023. The author V.A. is a recipient of the Marietta Blau-Grant (2023). The completion of this paper was supported and funded within the Post-DocTrack Program of the OeAW (85184), awarded to V.A. in 2025. B.F. is a recipient of the DOC Fellowship of the Austrian Academy of Sciences.

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 (accessed on 19 November 2024); https://doi.org/10.1016/j.jasrep.2021.102958 (accessed on 19 November 2024) or on request from the authors.

Acknowledgments

We extend our sincere gratitude to the Ministry of Culture of Bulgaria for granting permission for sampling and exportation. This work was conducted in collaboration with and under the supervision of the National Archaeological Institute with the Museum of the Bulgarian Academy of Sciences (NAIM-BAS), with special thanks to Krassimira Karadimitrova for her assistance during the sampling process in Sofia. We are particularly thankful to the director, Hristo Popov, for his invaluable support and guidance throughout this endeavour. We are indebted to the colleagues at the Regional Archaeological Museum in Plovdiv, Maya Martinova and Lyubomir Merdzhanov, for their collaboration during the sampling campaign. Furthermore, we would like to express our gratitude to Plamena Dakasheva, Wioletta Tenczar, and Murat Dirican for their expertise and assistance with the archaeological study, sample preparation, and laboratory work. We extend our thanks to Barbara Umfahrer for her meticulous organisation of the database. We are deeply grateful to Alexander Sokolicek (University of Salzburg) and Ben Russell (University of Edinburgh) for their invaluable guidance during the interpretation of the results. Lastly, we thank Emmanouil Anevlavis for his assistance in editing the photographs. We thank the anonymous reviewers for the fruitful feedback. We dedicate this publication to the memory of our director and group leader, Sabine Ladstätter, who passed away in June 2024. Her unwavering support and guidance were instrumental to this research, and we remain deeply grateful for her legacy and inspiration. This study derives from the PhD research conducted by V.A. at the University of Salzburg, under the supervision of Professor Alexander Sokolicek. The processes of material sampling, cataloguing, and photographing were integral components of the research.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Pedley, J. Sanctuaries and the Sacred in the Ancient Greek World; Cambridge University Press: Cambridge, UK, 2005. [Google Scholar]
  2. Scott, M. Space and Society in the Greek and Roman Worlds; Cambridge University Press: Cambridge, UK, 2012. [Google Scholar]
  3. Dobruski, V. Materiali po arheologijata na Bulgarija. SbNUNK 1896, 13, 398–442. [Google Scholar]
  4. Gocheva, Z. Trakijsko svetilishhe na Nimfite pri s. Ognjanovo. In Studia Archaeologica. Suppl. II. Sbornik Izsledvanija v Chest na Prof. Atanas Milchev; UI “Sv. Kliment Ohridski”: Sofia, Bulgaria, 2002; pp. 112–118. [Google Scholar]
  5. Karadimitrova, K. Pametnici na kulta na trite Nimfi ot NAM. GNAM 1992, 8, 189–195. [Google Scholar]
  6. Andreeva, P.; Anevlavi, V.; Frerix, B.; Cenati, C.; Prochaska, W.; Ladstätter, S.; Karadimitrova, K. A contribution to the urbanization and marble supply of Roman Thrace: An interdisciplinary study. Archaeol. Anthropol. Sci. 2024, 16, 23. [Google Scholar] [CrossRef]
  7. Anevlavi, V.; Prochaska, W.; Cenati, C.; Ivanov, I.; Ladstätter, S.; Popov, H.; Georgiev, P.; Kabakchieva, G. Geochemical and petrographic investigation of the provenance of white marble decorative elements from the Roman Villa Armira in south-eastern Bulgaria. Archaeol. Anthropol. Sci. 2022, 14, 229. [Google Scholar] [CrossRef]
  8. Anevlavi, V.; Prochaska, W.; Andreeva, P.; Ladstätter, S. New Evidence of the Supra-regional Marble Trade Network in Thrace, through the Archaeometric Study of Sculptures in Roman Philippopolis. J. Archaeol. Sci. 2025, 174, 106128. [Google Scholar] [CrossRef]
  9. 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]
  10. 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]
  11. 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, 45, 103582. [Google Scholar] [CrossRef]
  12. Prochaska, W.; Ladstätter, S.; Anevlavi, V. The challenge of a successful discrimination of ancient marbles (part IV): A databank for the white marbles from the region of Ephesos. J. Archaeol. Sci. Rep. 2024, 53, 104336. [Google Scholar] [CrossRef]
  13. Kovachev, V. (Ed.) Gold, Copper Mining and Geoarchaeology in Central Bulgaria. ACTA Mineralogica-Petrographica, Field Guide Series; Department of Mineralogy, Geochemistry and Petrology, University of Szeged: Szeged, Hungary, 2010; Volume 5. [Google Scholar]
  14. Zagorchev, I. Introduction to the Geology of SW Bulgaria. Geol. Balc. 2001, 31, 3–52. [Google Scholar] [CrossRef]
  15. Valchev, I. Izvangradskite Svetilishta V Rimskata Provintsiya Trakiya (I–IV Vek); Universitetsko Izdatelstvo “Sv. Kliment Ohridski”: Sofia, Bulgaria, 2015. [Google Scholar]
  16. Koleva, M. Rimska Idealna Skulptura ot Balgariya; Bulgarian Academy of Sciences: Sofia, Bulgaria, 2017. [Google Scholar]
  17. Gerasimov, T. Beležzki vărhu Pitiĭskite, Aleksandriĭskite i Kendri-zĭskite igri vaskite igri va\u030vr Filipopolis. In V Izsledvaniya v chest na akad Dimitŭr Dechev; Bulgarian Academy of Sciences: Sofia, Bulgaria, 1958; pp. 289–304. [Google Scholar]
  18. Sharankov, N. The Thracian κοινόν: New Epigraphic Evidence. In Thrace in the Graeco-Roman World. In Proceedings of the 10th International Congress of Thracology, Komotini and Alexandroupolis, Greece, 18–23 October 2005; Iakovidou, A., Ed.; National Hellenic Research Foundation: Athens, Greece, 2007; pp. 518–538. [Google Scholar]
  19. Boteva, D. Further considerations on the votive reliefs of the Thracian Horseman. In Moesica et Christiana. Studies in Honour of Professor Alexandru Barnea; Panaite, A., Cârjan, R., Căpiţă, C., Eds.; Muzeul Brăilei “Carol I”: Brăila, Romania, 2016; pp. 309–319. [Google Scholar]
  20. Boteva, D. Monuments and nature of the so-called Thracian Horseman again and not for the last time. In Συμπόσιον: Studies in Memory of Prof. Dimitar Popov; Delev, P., Ed.; St. Kliment Ohridski University Press: Sofia, Bulgaria, 2016; pp. 91–105. [Google Scholar]
  21. Ivanov, R. (Ed.) Tabula Imperii Romani. K35/2, Philippopolis; Tendril Publishing House: Sofia, Bulgaria, 2012. [Google Scholar]
  22. Tacheva-Hitova, M. Istorija na Iztochnite Kultove v Dolna Mizija i Trakija (VI v. pr. n. e.—IV v. ot. n. e.); Nauka i Izkustvo: Sofia, Bulgaria, 1982. [Google Scholar]
  23. Katsarov, G.I. Die Denkmäler des Thrakischen Reitergottes in Bulgarien; Karl W. Hiersemann: Leipzig, Germany, 1938. [Google Scholar]
  24. Velkov, I. Antični Pametnici iz Bǎlgarija. Godiš. Narodn. Muz. 1922, 1921, 198–216. [Google Scholar]
  25. Rabadjiev, K. The Triumph of Dionysos (and Herakles?). In Studia in Honorem Christo Danov, Thracia 12; Serdicae: Sofia, Bulgaria, 1998; pp. 201–206. [Google Scholar]
  26. Georgiev, P. Deo Soli Invicto Mithrae—Mithra’s Invincible Sun God; Sema RSH: Sofia, Bulgaria, 2021; ISBN 978-954-9670-52-3. [Google Scholar]
  27. Topalilov, I. Philippopolis. The city from the 1st to the beginning of the 7th century. In Roman Cities in Bulgaria; Ivanov, R., Ed.; National Archaeological Institute with Museum: Sofia, Bulgaria, 2012; Volume 1, pp. 363–437. [Google Scholar]
  28. Mihailov, G. (Ed.) Inscriptiones Graecae in Bulgaria Repertae, Vol. III,1. Inscriptiones inter Haemum et Rhodopem Repertae. In Fasciculus Prior: Territorium Philippopolis; (IGBulg III,1); RIVA: Sofia, Bulgaria, 1961. [Google Scholar]
  29. Cherneva, S. Unfinished Marble Monuments from Philippopolis: An evidence of local production. Annu. Reg. Archaeol. Mus. 2009, 11, 130–137. [Google Scholar]
  30. Dilaria, S.; Bonetto, J.; Germinario, L.; Previato, C.; Girotto, C.; Mazzoli, C. The stone artifacts of the National Archaeological Museum of Adria (Rovigo, Italy): A noteworthy example of heterogeneity. Archaeol. Anthropol. Sci. 2024, 16, 14. [Google Scholar] [CrossRef]
  31. Diffendale, D.P.; Marra, F.; Gaeta, M.; Terrenato, N. Combining geochemistry and petrography to provenance Lionato and Lapis Albanus tuffs used in Roman temples at Sant’Omobono, Rome, Italy. Geoarchaeology 2019, 34, 187–199. [Google Scholar] [CrossRef]
  32. Gluhak, T.M.; Hofmeister, W. Geochemical provenance analyses of Roman lava millstones north of the Alps: A study of their distribution and implications for the beginning of Roman lava quarrying in the Eifel region (Germany). J. Archaeol. Sci. 2011, 38, 1603–1620. [Google Scholar] [CrossRef]
  33. Dilaria, S.; Ricci, G.; Secco, M.; Beltrame, C.; Costa, E.; Giovanardi, T.; Bonetto, J.; Artioli, G. Vitruvian binders in Venice: First evidence of Phlegraean pozzolans in an underwater Roman construction in the Venice Lagoon. PLoS ONE 2024, 19, e0313917. [Google Scholar] [CrossRef]
  34. Lancaster, L.; Sottili, G.; Marra, F.; Ventura, G. Provenancing of lightweight volcanic stones used in ancient Roman concrete vaulting: Evidence from Rome. Archaeometry 2011, 53, 707–727. [Google Scholar] [CrossRef]
  35. Poretti, G.; Brilli, M.; De Vito, C.; Conte, A.M.; Borghi, A.; Günther, D.; Zanetti, A. New considerations on trace elements for quarry provenance investigation of ancient white marbles. J. Cult. Herit. 2017, 28, 16–26. [Google Scholar] [CrossRef]
  36. Vinci, M.S. Ancient stone quarries on the Iberian peninsula: Updated research and methodological approaches—AG Gutiérrez Garcia-Moreno, and P. Rouillard, eds. 2018. Lapidum natura restat: Canteras antiguas de la península ibérica en su contexto (cronología, técnicas y organización de la explotación); Institut Català d’Arqueologia Clàssica and Casa de Velázquez: Tarragona and Madrid, Spain, 2018; pp. 197. J. Rom. Archaeol. 2021, 34, 499–505. [Google Scholar]
  37. Ruppiene, V.; Schüssler, U.; Unterwurzacher, M. Auerbach Marble Quarries in the Odenwald near Hochstädten. Min. Cult. Landsc. IES Yearb. 2013, 120–129. [Google Scholar]
  38. Gutiérrez Garcia-M, A.; Royo Plumed, H.; González Soutelo, S.; Savin, M.C.; Lapuente, P.; Chapoulie, R. The marble of O Incio (Galicia, Spain): Quarries and first archaeometric characterisation of a material used since Roman times. Archéosciences Rev. Archéom. 2016, 40, 103–177. [Google Scholar] [CrossRef]
  39. Antonelli, F.; Lazzarini, L. An updated petrographic and isotopic reference database for white marbles used in antiquity. Rend. Lincei Sci. Fis. Nat. 2015, 26, 399–413. [Google Scholar] [CrossRef]
  40. Bruno, M.; Elçi, H.; Yavuz, A.B.; Attanasio, D. Unknown ancient marble quarries of Western Asia Minor. In ASMOSIA IX. Interdisciplinary Studies on Ancient Stone. In Proceedings of the IX ASMOSIA Conference, Tarragona, Spain, 8–13 June 2009; Gutiérrez Garcia-M, A., Lapuente, P., Rodà, I., Eds.; Institut Català d’Arqueologia Clàssica: Tarragona, Spain, 2012; pp. 562–572. [Google Scholar]
  41. Wielgosz-Rondolino, D.; Antonelli, F.; Bojanowski, M.J.; Gładki, M.; Göncüoğlu, M.C.; Lazzarini, L. Improved methodology for identification of Göktepe white marble and the understanding of its use: A comparison with Carrara marble. J. Archaeol. Sci. 2020, 113, 105059. [Google Scholar] [CrossRef]
  42. Wielgosz-Rondolino, D.; Antonelli, F.; Bojanowski, M.J.; Degryse, P.; Gładki, M.; Göncüoğlu, M.C.; Lazzarini, L.; Long, L.; Mandera, S. Ancient white and grey marbles from Aphrodisias: Multi-method characterization. Archaeometry 2024, 66, 30–56. [Google Scholar] [CrossRef]
  43. Anevlavi, V.; Prochaska, W.; Dimitrov, Z.; Ladstätter, S. Provenance Matters. A Multi-Proxy Approach for the Determination of White Marbles from Villa Armira, Bulgaria. In Proceedings of the ASMOSIA XII—Association for the Study of Marble & Other Stones in Antiquity XII, Izmir, Turkey, 8–14 October 2018; International Conference Proceedings. Yavuz, A.B., Yolaçan, B., Bruno, M., Eds.; Publisher: Dokuz Eylül University Publications: Izmir, Turkey, 2023; pp. 165–174. [Google Scholar]
  44. Blackburn, W.H.; Dennen, W.H. Principles of Mineralogy; Wm. C. Brown Publishers: Dubuque, IA, USA, 1994. [Google Scholar]
  45. Lepsius, G.R. Griechische Marmorstudien; Verlag der Königlichen Akademie der Wissenschaften: Berlin, Germany, 1890. [Google Scholar]
  46. Moltesen, M. The Lepsius Marble Samples; Ny Carlsberg Glyptotek: Copenhagen, Denmark, 1994. [Google Scholar]
  47. Craig, H.; Craig, V. Greek marbles: Determination of provenance by isotopic analysis. Science 1972, 176, 401–403. [Google Scholar] [CrossRef]
  48. Bowen, R. Isotopes in the Earth Sciences; Elsevier Applied Science: London, UK, 1988. [Google Scholar]
  49. Attanasio, D.; Brilli, M.; Ogle, N. The Isotopic Signature of Classical Marbles; L’Erma di Bretschneider: Rome, Italy, 2006. [Google Scholar]
  50. Coleman, M.; Walker, S. Stable isotope identification of Greek and Turkish marbles. Archaeometry 1979, 21, 107–112. [Google Scholar] [CrossRef]
  51. Polikreti, K. Detection of ancient marble forgery: Techniques and limitations. Archaeometry 2007, 49, 603–619. [Google Scholar] [CrossRef]
  52. Poretti, G. In Situ Analysis of White Marble from the Mediterranean Basin by LA-ICP-MS: Inferences on Provenance Based on Trace-Element Profiles. Ph.D. Thesis, Sapienza—University of Rome, Rome, Italy, 24 October 2016. [Google Scholar]
  53. Anevlavi, V.; Andreeva, P.; Prochaska, W.; Karadimitrova, K. Cults and Materials: Provenance Investigations of the White Marble Votive Plaques from the Sanctuary of Apollo Γεικεσηνος in Roman Thrace. In Proceedings of the Workshop: “Danubian Riders”—Deconstruction of a Religious Assemblage; Kremer, Ed.; Vienna, forthcoming.
  54. Koleva, M. The Visual Language of Roman Art and Roman Thrace. In Izkustvovedski Cheteniya 2016; Tekstove, nadpisi, obrazi; Moutafov, E., Erdeljan, J., Eds.; Bulgarian Academy of Sciences: Sofia, Bulgaria, 2017; pp. 17–27. [Google Scholar]
  55. Prochaska, W.; Anevlavi, V.; Chakarov, K.; Andreeva, P.; Sutev, I. Archaeometric Investigations on the Significance of Limestones in the Roman Provinces of the Southern Lower Danube. Minerals 2025, 15, 267. [Google Scholar] [CrossRef]
  56. Andreeva, P.; Anevlavi, V.; Prochaska, W.; Ladstätter, S. Statuary in the marble trade networks of Roman Thrace: A case study of unfinished and replicated products. In Proceedings of the ASMOSIA XIII Conference, Vienna, Austria, 19–23 September 2022; Ladstätter, S., Prochaska, W., Anevlavi, V., Eds.; Sonderschriften des Österreichischen Archäologischen Instituts: Vienna, Austria, 2025; pp. 331–349. [Google Scholar]
  57. Zuglev, K. Antichni Arheologicheski Nahodki ot Asenovgradsko. Archaeologia 1965, 4, 38–43. [Google Scholar]
  58. Madzharov, M. Rimski patishta v Balgariya. Prinos v razvitieto na rimskata patna sistema v provintsiite Miziya i Trakiya. Roman Roads in Bulgaria: Contribution to the Development of Roman Road System in the Provinces of Moesia and Thrace; Faber: Veliko Tarnovo, Bulgaria, 2009. [Google Scholar]
  59. Kolb, A. Transport in Thracia. In Proceedings of the First International Roman and Late Antique Thrace Conference: Cities, Territories and Identities, Plovdiv, Bulgaria, 3–7 October 2016; Vagalinski, L., Raycheva, M., Boteva, D., Sharankov, N., Eds.; Izvestiya na Natsionalniya arheologicheski institut 44; Bulgarian Academy of Sciences: Sofia, Bulgaria, 2018; pp. 269–277. [Google Scholar]
  60. Sayar, M.H. Perinthos-Herakleia (Marmara Ereğlisi) und Umgebung. Geschichte, Testimonien, griechische und lateinische Inschriften; Denkschriften der philosophisch-historischen Klasse 269; Verlag der Österreichischen Akademie der Wissenschaften: Wien, Austria, 1998. [Google Scholar]
  61. Raepsaet, G. Land Transport, Part 2: Riding, Harnesses, and Vehicles. In The Oxford Handbook of Engineering and Technology in the Classical World; Oleson, J.P., Ed.; Oxford University Press: Oxford, UK, 2008; pp. 580–605. [Google Scholar]
  62. Bernard, S.G. The Transport of Heavy Loads in Antiquity: Lifting, Moving, and Building in Ancient Rome. In Perspektiven der Spolienforschung 1. Spoliierung und Transposition; Altekamp, S., Marcks-Jacobs, C., Seiler, P., Eds.; Topoi 15; Edition Topoi: Berlin, Germany, 2013; pp. 99–122. [Google Scholar]
  63. Riemer, H.; Förster, F. Donkeys, Camels, and the Logistics of Ancient Caravan Transport. In Caravans in Socio-Cultural Perspective. Past and Present; Clarkson, P.B., Santoro, C.M., Eds.; Routledge: London, UK, 2022; pp. 145–173. [Google Scholar]
Figure 1. (a) Map of the Roman province of Thrace in the mid-2nd century AD, showing its territorial extent and neighbouring provinces. To the northwest lies Moesia Superior, while Moesia Inferior borders the province along the Danube to the north and northeast. The location of Ognyanovo village is indicated in red. (B. Frerix, ÖAI/ÖAW). (b) Map of Roman Thrace focusing on the region of Philippopolis. Triangular markers indicate sanctuaries. For more information on the sanctuaries, see Table S2. (Map: B. Frerix, ÖAI/ÖAW).
Figure 1. (a) Map of the Roman province of Thrace in the mid-2nd century AD, showing its territorial extent and neighbouring provinces. To the northwest lies Moesia Superior, while Moesia Inferior borders the province along the Danube to the north and northeast. The location of Ognyanovo village is indicated in red. (B. Frerix, ÖAI/ÖAW). (b) Map of Roman Thrace focusing on the region of Philippopolis. Triangular markers indicate sanctuaries. For more information on the sanctuaries, see Table S2. (Map: B. Frerix, ÖAI/ÖAW).
Minerals 15 00670 g001aMinerals 15 00670 g001b
Figure 2. Polarised (a) and transmitted (b) light photos of votive plate FWM0047 (calcite), exhibiting a fine-to-medium-grained ultra-fine-mylonitic matrix with a mosaic fabric and straight-to-curved grain boundaries. The sample contains opaque minerals, including iron oxide (possibly hematite), and displays a blastomylonitic groundmass with deformed relics surrounded by corrosion features—indicative of high-strain deformation, characteristic of Asenovgrad marble (ph. V. Anevlavi, ÖAI/ÖAW).
Figure 2. Polarised (a) and transmitted (b) light photos of votive plate FWM0047 (calcite), exhibiting a fine-to-medium-grained ultra-fine-mylonitic matrix with a mosaic fabric and straight-to-curved grain boundaries. The sample contains opaque minerals, including iron oxide (possibly hematite), and displays a blastomylonitic groundmass with deformed relics surrounded by corrosion features—indicative of high-strain deformation, characteristic of Asenovgrad marble (ph. V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g002
Figure 3. Polarised (a) and transmitted (b) light photos of votive plate FWM0104 (calcite), displaying a fine-grained mosaic texture with straight-to-curved grain boundaries and an ultra-fine-mylonitic matrix (blastomylonitic groundmass). The sample includes opaque minerals, sheet silicates (mica, chlorite), apatite, and exhibits undulose extinction in larger crystals—indicative of dynamic deformation typical of marbles from the Asenovgrad region (ph. V. Anevlavi, ÖAI/ÖAW).
Figure 3. Polarised (a) and transmitted (b) light photos of votive plate FWM0104 (calcite), displaying a fine-grained mosaic texture with straight-to-curved grain boundaries and an ultra-fine-mylonitic matrix (blastomylonitic groundmass). The sample includes opaque minerals, sheet silicates (mica, chlorite), apatite, and exhibits undulose extinction in larger crystals—indicative of dynamic deformation typical of marbles from the Asenovgrad region (ph. V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g003
Figure 4. Polarised (a) and transmitted (b) light of votive plate FWM0353 (calcite), featuring an ultra-fine-mylonitic matrix (blastomylonitic groundmass) with a mortar texture and sutured-grain boundaries. The sample shows slightly deformed twin lamellae, undulose extinction, and mica and opaque mineral inclusions—hallmarks of a high-strain metamorphic fabric consistent with Asenovgrad marble (ph. V. Anevlavi, ÖAI/ÖAW).
Figure 4. Polarised (a) and transmitted (b) light of votive plate FWM0353 (calcite), featuring an ultra-fine-mylonitic matrix (blastomylonitic groundmass) with a mortar texture and sutured-grain boundaries. The sample shows slightly deformed twin lamellae, undulose extinction, and mica and opaque mineral inclusions—hallmarks of a high-strain metamorphic fabric consistent with Asenovgrad marble (ph. V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g004aMinerals 15 00670 g004b
Figure 5. Polarised (a) and transmitted (b) light of marble sample FWM0846 from the Asenovgrad quarry, showing a mortar texture with sutured-grain boundaries. The sample features opaque minerals (possibly pyrite), deformed twin lamellae, and a cataclastic texture. It displays alternating layers of fine-grained matrices and larger, corroded crystals, indicating mild deformation—consistent with the geological characteristics of Asenovgrad marble and the archaeological samples presented above (ph. V. Anevlavi, ÖAI/ÖAW).
Figure 5. Polarised (a) and transmitted (b) light of marble sample FWM0846 from the Asenovgrad quarry, showing a mortar texture with sutured-grain boundaries. The sample features opaque minerals (possibly pyrite), deformed twin lamellae, and a cataclastic texture. It displays alternating layers of fine-grained matrices and larger, corroded crystals, indicating mild deformation—consistent with the geological characteristics of Asenovgrad marble and the archaeological samples presented above (ph. V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g005
Figure 6. Isotopic diagram of the sampled artefacts, showing the group of the votive reliefs. Red squares represent votive reliefs from the Sanctuary of the Three Nymphs; blue circles denote other votive reliefs. Grey squares indicate samples from SE Rhodope, magenta triangles correspond to Asenovgrad marble, and black circles represent samples from Prokonnesos. (ed. by V. Anevlavi, ÖAI/ÖAW).
Figure 6. Isotopic diagram of the sampled artefacts, showing the group of the votive reliefs. Red squares represent votive reliefs from the Sanctuary of the Three Nymphs; blue circles denote other votive reliefs. Grey squares indicate samples from SE Rhodope, magenta triangles correspond to Asenovgrad marble, and black circles represent samples from Prokonnesos. (ed. by V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g006
Figure 7. Multivariate diagram of the sampled artefacts of this group. Red squares represent votive reliefs from the Sanctuary of the Three Nymphs; blue circles denote other votive reliefs. Grey squares indicate samples from SE Rhodope, magenta triangles correspond to Asenovgrad marble, and black circles represent samples from Prokonnesos (ed. by V. Anevlavi, ÖAI/ÖAW).
Figure 7. Multivariate diagram of the sampled artefacts of this group. Red squares represent votive reliefs from the Sanctuary of the Three Nymphs; blue circles denote other votive reliefs. Grey squares indicate samples from SE Rhodope, magenta triangles correspond to Asenovgrad marble, and black circles represent samples from Prokonnesos (ed. by V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g007
Figure 8. Graphic presentation of the sampled elements of the votive reliefs from the Sanctuary of the Three Nymphs and their provenance (ed. by V. Anevlavi, ÖAI/ÖAW).
Figure 8. Graphic presentation of the sampled elements of the votive reliefs from the Sanctuary of the Three Nymphs and their provenance (ed. by V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g008
Figure 9. Graphic presentation of the sampled elements of the votive reliefs and their provenance (ed. by V. Anevlavi, ÖAI/ÖAW).
Figure 9. Graphic presentation of the sampled elements of the votive reliefs and their provenance (ed. by V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g009
Figure 10. The chronological distribution of the sampled votive reliefs (ed. by V. Anevlavi, ÖAI/ÖAW).
Figure 10. The chronological distribution of the sampled votive reliefs (ed. by V. Anevlavi, ÖAI/ÖAW).
Minerals 15 00670 g010
Table 1. The sampled artefacts from the Sanctuary of the three Nymphs.
Table 1. The sampled artefacts from the Sanctuary of the three Nymphs.
Project NrLab NrM. Nr *ObjectMaterialStone DescriptionDatingLocationFind SpotDetails on Finding-SpotProvenance ResultsPhotoBibliography
FWM006292491288Votive reliefMarbleWhite colour; coarse-grained.Mid-3rd century ADNAIM MuseumPhilippopolis; Regionn/aAsenovgradMinerals 15 00670 i001[5], pp. 189–195
Description: The three nymphs are holding hands under an arcade.
FWM00999281962Votive plateMarbleWhite colour; fine-grained with larger crystals.Mid-3rd century ADNAIM MuseumOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i002[5], pp. 189–195
Description: Relief images of three nymphs. Parts of the lower frame are missing.
FWM01009282n/aVotive plateMarbleWhite colour; medium-grained with larger crystals.Mid-3rd century ADNAIM MuseumOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i003[5], pp. 189–195
Description: Relief images of three nymphs and Apollo holding a patera over an altar.
FWM01509283n/aVotive plate MarbleWhite colour; fine-grained with larger crystals.2nd–3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i004IGBulg III,1 1357
Description: Relief images of three nymphs. Greek inscription engraved on the frame.
FWM016492841076Votive plateMarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i005IGBulg III,1 1357
Description: Relief images of three nymphs. Greek inscription engraved on the frame.
FWM016592852745Votive plateMarbleWhite colour; fine-grained with larger crystals.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of
the three Nymphs
n/aAsenovgradMinerals 15 00670 i006[5], pp. 189–195
Description: Relief images of three nymphs holding hands.
FWM01669286969Votive plateMarbleWhite colour; medium-to-coarse-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i007[5], pp. 189–195
Description: Relief images of three dancing nymphs. Greek inscription engraved on the frame.
FWM01679287965Votive plate MarbleWhite colour; medium-to-coarse-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i008[5], pp. 189–195
Description: Relief images of three nymphs. Greek inscription engraved on the lower frame.
FWM01689288994Votive plate MarbleWhite colour; medium-to-coarse-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i009[5], pp. 189–195
Description: Relief images of three nymphs holding hands.
FWM01699289964Votive plateMarbleWhite colour; very fine-grained; ultra mylonite.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i010[5], pp. 189–195
Description: Relief images of three nymphs.
FWM01709290974Votive plate MarbleWhite colour; very fine-grained; brown spots; fine mylonite.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i011[5], pp. 189–195
Description: Relief images of three nymphs. Greek inscription engraved on the frame.
FWM034797071014Votive plate MarbleWhite colour; fine-grained with larger crystals.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i012[5], pp. 189–195
Description: Relief images of three nymphs holding hands.
FWM034897081008Votive plate MarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aProkonnesosMinerals 15 00670 i013[5], pp. 189–195
Description: Relief images of three nymphs holding hands; only a fragment showing the body of the rightmost nymph is preserved.
FWM034997091010Votive plateMarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i014[5], pp. 189–195
Description: Relief images of three nymphs.
FWM035097101009Votive plate MarbleWhite colour; fine-grained with larger crystals.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i015[5], pp. 189–195
Description: Relief images of three nymphs holding hands.
FWM035197111001Votive plate MarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i016[5], pp. 189–195
Description: Relief images of three nymphs.
FWM035297121006Votive plateMarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i017[5], pp. 189–195
Description: Relief images of three nymphs.
FWM035397131007Votive plate MarbleWhite colour; medium-grained.Mid-3rd century ADNAIM DepotOgnyanovo

Sanctuary of the three Nymphs
n/aAsenovgradMinerals 15 00670 i018[5], pp. 189–195
Description: Relief images of three nymphs holding hands.
* Museum number.
Table 3. The calculated statistical parameters of the investigated marble samples. Variables: Mn, Mg, Fe, Sr, Cr, V, Ba, Y, Cd, La, Ce, Yb, U, δ18O‰, and δ13C‰ (ed. by V. Anevlavi, ÖAI/ÖAW).
Table 3. The calculated statistical parameters of the investigated marble samples. Variables: Mn, Mg, Fe, Sr, Cr, V, Ba, Y, Cd, La, Ce, Yb, U, δ18O‰, and δ13C‰ (ed. by V. Anevlavi, ÖAI/ÖAW).
SampleDistanceAbs. Prob.Rel. Prob.ProvenanceRel. Prob.Provenance
1. choice2. choice
Sanctuary of the Nymphs
9249FWM00624.410.899.2Asenovgrad0.8Prokonnesos
9281FWM00991.157.4100.0Asenovgrad0.0SE Rhodope
9282FWM01003.021.4100.0Asenovgrad0.0Prokonnesos
9283FWM01501.839.2100.0Asenovgrad0.0Prokonnesos
9284FWM01640.194.2100.0Asenovgrad0.0Prokonnesos
9285FWM01654.88.9100.0Asenovgrad0.0SE Rhodope
9286FWM01665.08.2100.0Asenovgrad0.0Prokonnesos
9287FWM01670.194.6100.0Asenovgrad0.0SE Rhodope
9288FWM01686.14.6100.0Asenovgrad0.0SE Rhodope
9289FWM01690.672.8100.0Asenovgrad0.0SE Rhodope
9290FWM01703.814.999.9Asenovgrad0.1Prokonnesos
9707FWM03471.839.3100.0Asenovgrad0.0Prokonnesos
9708FWM03481.351.1100.0Prokonnesos0.0Asenovgrad
9709FWM03491.350.3100.0Asenovgrad0.0SE Rhodope
9710FWM03502.922.4100.0Asenovgrad0.0Prokonnesos
9711FWM03510.866.9100.0Asenovgrad0.0Prokonnesos
9712FWM03524.98.4100.0Asenovgrad0.0Prokonnesos
9713FWM03534.510.299.9Asenovgrad0.1Prokonnesos
Miscellaneous Votive Reliefs
9243FWM00473.318.999.8Asenovgrad0.2Prokonnesos
9244FWM00493.715.199.9Asenovgrad0.1Prokonnesos
9264FWM01089.01.184.8Asenovgrad10.0Prokonnesos
9265FWM01121.253.1100.0Asenovgrad0.0Prokonnesos
9267FWM01021.936.8100.0Asenovgrad0.0Prokonnesos
9268FWM01044.98.399.1Asenovgrad0.9Prokonnesos
9269FWM01051.060.4100.0Asenovgrad0.0SE Rhodope
9815FWM04605.46.597.2Asenovgrad2.8Prokonnesos
9816FWM04617.42.499.9Asenovgrad0.1Prokonnesos
9817FWM046276.60.0100.0Unknown
Asenovgrad *
0.0SE Rhodope
9827FWM04723.021.499.9Asenovgrad0.1Prokonnesos
9829FWM04741.059.1100.0Asenovgrad0.0Prokonnesos
9831FWM04765.46.5100.0Asenovgrad0.0Prokonnesos
9837FWM04822.725.6100.0Asenovgrad0.0SE Rhodope
9838FWM04836.83.292.8Asenovgrad7.2Prokonnesos
9839FWM04840.962.6100.0Asenovgrad0.0SE Rhodope
9840FWM04855.27.499.6Asenovgrad0.4Prokonnesos
9841FWM04861.059.0100.0Asenovgrad0.0Prokonnesos
9842FWM04873.913.799.6Asenovgrad0.4Prokonnesos
9833FWM04781.297.299.3Asenovgrad0.7SE Rhodope
9241FWM00411.195.296.0SE Rhodope2.9Asenovgrad
9832FWM04772.084.699.2Prokonnesos0.7Dokimeion
9859FWM05040.199.997.9Asenovgrad2.0SE Rhodope
9860FWM05056.029.977.2Asenovgrad15.9Dokimeion
* The sample shares identical geochemical data with the Asenovgrad marble source, yet a minority displays partial divergence. This variance might be attributed to the potential presence of an undiscovered source, such as a quarry or outcrop, in the vicinity of Asenovgrad.
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

Anevlavi, V.; Prochaska, W.; Andreeva, P.; Petkova, K.; Frerix, B. White Marble Sourcing and Regional Workshop Dynamics in Roman Thrace: An Archaeometric Study of Votive Reliefs. Minerals 2025, 15, 670. https://doi.org/10.3390/min15070670

AMA Style

Anevlavi V, Prochaska W, Andreeva P, Petkova K, Frerix B. White Marble Sourcing and Regional Workshop Dynamics in Roman Thrace: An Archaeometric Study of Votive Reliefs. Minerals. 2025; 15(7):670. https://doi.org/10.3390/min15070670

Chicago/Turabian Style

Anevlavi, Vasiliki, Walter Prochaska, Petya Andreeva, Kalina Petkova, and Benjamin Frerix. 2025. "White Marble Sourcing and Regional Workshop Dynamics in Roman Thrace: An Archaeometric Study of Votive Reliefs" Minerals 15, no. 7: 670. https://doi.org/10.3390/min15070670

APA Style

Anevlavi, V., Prochaska, W., Andreeva, P., Petkova, K., & Frerix, B. (2025). White Marble Sourcing and Regional Workshop Dynamics in Roman Thrace: An Archaeometric Study of Votive Reliefs. Minerals, 15(7), 670. https://doi.org/10.3390/min15070670

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

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.
Back to TopTop