Cinnabar for Roman Ephesus: Material quality, processing and provenance

Page 1 of 23 CINNABAR FOR ROMAN EPHESUS: MATERIAL QUALITY, 1 PROCESSING AND PROVENANCE 2 3 Alexandra RODLER-RØRBO, 1,2,3,4* Anthony J. BARAGONA, 5 Eliah J. VERBEEMEN, 1 Lasse Vilien 4 SØRENSEN, 6 Berk ÇAKMAKOĞLU, 7 Cahit HELVACI, 7 Eduardo BOLEA-FERNANDEZ, 8,9 Ana 5 RUA-IBARZ, 8,9 Frank VANHAECKE, 8 Hilary BECKER, 10 Gilberto ARTIOLI, 11 Lilli ZABRANA, 3 6 Vinciane DEBAILLE, 2 Nadine MATTIELLI, 2 Steven GODERIS, 1 Philippe CLAEYS 1 7 8 1 Archaeology, Environmental Changes and Geo-Chemistry, Vrije Universiteit Brussel, VUB, Pleinlaan 9 2, 1050 Brussels, Belgium 10 2 Laboratoire G-time, Department of Geosciences, Environment and Society, Université Libre de 11 Bruxelles, ULB, Avenue F.D. Roosevelt 50, 1050 Brussels, Belgium 12 3 Austrian Archaeological Institute of the Austrian Academy of Sciences, Dominikanerbastei 16, 1010 13 Vienna, Austria 14 4 Human Evolution and Archaeological Sciences (HEAS), University of Vienna, Austria 15 5 Freelance, Associated with the University of Applied Arts, Vienna, Salzgries 14, 1090 Vienna, Austria 16 6 National Museum of Denmark, Frederiksholms Kanal 12, 1220 Copenhagen, Denmark 17 7 Department of Geology Engineering, Dokuz Eylül University, Tınaztepe Yerleşkesi, İzmir, Turkey 18 8 Atomic and Mass Spectrometry – A&MS research unit, Department of Chemistry, Ghent University, 19 Campus Sterre, Krijgslaan 281-S12, 9000 Ghent, Belgium 20 9 Aragón Institute of Engineering Research (I3A), Department of Analytical Chemistry, University of 21 Zaragoza, Zaragoza, Spain 22 10 Department of Middle Eastern and Ancient Mediterranean Studies, Binghamton University. State 23 University of New York Binghamton, USA 24 11 Department of Geosciences, University of Padua, Via Gradenigo 6, I-35131 Padua, Italy 25 * Corresponding author: [email protected]; Current address: Austrian Archaeological 26 Institute of the Austrian Academy of Sciences, Dominikanerbastei 16, 1010 Vienna, Austria 27 28 29 Page 2 of 23 ABSTRACT 30 Ephesus was an important harbor city that flourished during the Roman period and ancient texts mention 31 Almadén in Spain and the Cilbian fields of Ephesus as important cinnabar sources in antiquity. This work 32 investigates whether imported cinnabar was used and whether this could be related to changes in painting 33 activities over time. Microscopic analysis indicates a consistent preparation of cinnabar, hinting at a 34 uniform source material quality or processing technique. However, the use of cinnabar varies among the 35 architectural structures studied, indicating a plurality of painting techniques. A few of the analyzed cinnabar 36 samples overlap with Turkishand Balkan reference Pb isotope ratios; three samples from tabernas, 37 however, deviate from this. The Hg isotope ratios reveal that cinnabar from carbonate-hosted deposits was 38 likely used, and that processing of cinnabar included heating as suggested by ancient texts. Most notably, 39 a correlation exists between the geochemical data and the painting technique – shifts in sourcing and 40 cinnabar usage are potentially assignable to building chronology and/or usage. Through the lens of material 41 provenance and processing, Ephesian cinnabar brings the organization of pigment trade into focus. 42 43 Keywords: Microscopic analysis, Pb isotope ratios, Hg isotope ratios, wall paintings, pigment trade 44 45 46 1. INTRODUCTION 47 Various types of mineral pigments were used in Roman art – some of these materials might have been 48 readily accessible, while others were most likely traded. Among the rarest and most highly valued materials 49 used as pigment in antiquity was cinnabar (α-HgS), which was known as μίνιον in Greek and minium in 50 Latin (inter al. Theophrastus De Lap. 58; Vitruvius De Arch. 7.8.1-7.9.6; Pliny HN 33.37-40; Dioscorides 51 MM 5.94 in Beck 2005, 374; Becker, 2022). These terms should not be confused with Indian cinnabar (or 52 dragon’s blood (Indica cinnabaris) and minium secundarium (or red lead, which was also known as cerussa 53 usta). Despite the widespread use of pigments throughout history (e.g., Hunt-Ortiz et al., 2011; Cooke et 54 al., 2013; Emslie et al., 2015) and the particularly high demand for various colorants during the Roman 55 period, as evidenced by the rich wall paintings found in both private and public buildings (e.g., Esposito, 56 2017; Slavdori and Sbrolli, 2021), there is limited knowledge about the organization of pigment 57 production/processing and trade. This is puzzling, considering that ancient authors such as Theophrastus 58 (c. 371–287 BCE) in De Lapidibus (De Lap.), Vitruvius (c. 80/70–15 BCE) in De Architectura (De arch.), 59 Pliny the Elder (23/24–79 CE) in Historia Naturalis (HN), and Dioscorides (c. 40–90 CE) in De Materia 60 Medica (MM), highlight the significance of provenance for the value of ancient pigments. The Cilbian fields 61 of Ephesus and the Almadén mines in Spain were mentioned as the most important cinnabar sources during 62 the Roman period (Gliozzo, 2021); whether Monte Amiata in Grosseto (Italy) or Idrija in Slovenia have 63 been exploited during the Roman period is not certain (Spangenberg et al., 2010). Furthermore, several 64 surveys have already investigated the geology and geochemistry of larger relevant ore bodies in Turkey 65 (e.g., Yıldız and Bailey, 1978; Gemici, 2008; Sayre et al., 2001) and mention for example the Tmolus Mt. 66 near Sardis (Foss and Hanfmann, 1975) or cinnabar outcrops near Selçuk (Schmiesser, 1906). Cinnabar can 67 Page 3 of 23 occur as a vein mineral, in strata-bound deposits such as in Almadén, Spain, or as secondary mineral in 68 gossans such as in Las Cruces, Spain (e.g., Higueras et al., 2005; Palero-Fernández et al., 2015; Yesares et 69 al., 2017). While ancient texts provide highly valuable clues for evaluating cinnabar provenance, 70 archaeometric analysis – drawing on geological, mineralogical-petrographic and geochemical data – can 71 provide crucial information on material provenance; together these approaches have the potential of 72 providing a way forward for understanding the organization of pigment trade and production networks. The 73 city of Ephesus (present-day Selçuk, south of İzmir, Turkey) developed near the mouth of the Küçük 74 Menderes River on the west coast of Anatolia (for a geochemical discussion of the Ephesian harbor, see 75 Delile et al., 2015). During the Hellenistic and Roman periods, Ephesus emerged as an important trading 76 hub of Asia Minor and the Aegean area. In 188 BCE, Ephesus became part of the Pergamene Empire, 77 leading to significant investments in the reorganization of the harbor and urbanization of Ephesus, which 78 were the foundations for its emergence as an exceptionally important and influential commercial center of 79 the ancient Mediterranean world and its success during the Roman period (Ladstätter, 2016). In 133 BCE, 80 Ephesus became part of the Roman Republic and the capital of the province Asia Minor. Already during 81 the 2nd c. BCE, local artisans began to produce ceramics such as table ware and lamps for exportation 82 (Ladstätter, 2016; Fragnoli et al., 2022). However, despite the significance of Ephesus in the production of 83 ceramic lamps during the Hellenistic period, pottery workshops from this period have not been found so far 84 (Fragnoli et al., 2022). Ephesus was a thriving trade center and a pigment consumer, which raises the 85 question whether this site only imported pigments, or whether processing also occurred here. Currently 86 there is no evidence of pigment production or processing workshops in or near Ephesus, however, this does 87 not necessarily mean that no (local) raw materials were processed for use as painting material. Only a few 88 pigment workshops are currently known for the Hellenistic and Roman periods and their trade networks 89 are still not fully understood. These include sites in the Bay of Naples, Italy (1st c. BCE; e.g., Lazzarini and 90 Verità, 2015), in Memphis, Egypt (1st c. BCE; e.g., Nicholson, 2003), and the late Hellenistic multi-crafting 91 workshop of Kos, Greece (Kostomitsopoulou Marketou, 2019). 92 Identifying the material quality, processing and provenance of cinnabar can therefore enhance our 93 understanding of ancient pigment workshop practices and the reconstruction of trade routes (e.g., Hunt-94 Ortiz et al., 2011; Prieto et al., 2015; Rodríguez et al., 2020; Minami et al., 2021). For example, recently, 95 the provenance of cinnabar used in wall paintings in the Roman city of Averticum (Avenches, Vaud, 96 Switzerland) has been investigated suggesting that cinnabar was possibly from the Almadén mining district 97 in Spain (Spangenberg et al., 2010). The analysis of pigment raw materials applies 98 mineralogical/petrographic and geochemical approaches to evaluate possible raw material sources (e.g., 99 Spangenberg et al., 2010; Hunt-Ortiz et al., 2011; Rodler et al., 2017). Geochemical approaches have 100 included lead (Pb) and/or sulphur (S) isotopic analysis as well as mercury (Hg) isotopic analysis (recently 101 reviewed by Gliozzo, 2021). The latter has been used for evaluating which cinnabar deposits were exploited 102 during Andean prehistory (Cooke et al., 2013). Pb is a trace element in sulphur-bearing minerals such as 103 cinnabar, and Pb isotopic analysis has long been used for the provenance study of various inorganic artifacts 104 (e.g., Stos-Gale and Gale, 2009; Artioli et al., 2016). Lead isotopic analysis makes use of the four isotopes 105 of lead 204Pb, 206Pb, 207Pb and 208Pb that have a natural relative abundance of ca. 1.4 %, 24.1 %, 22.1 % and 106 52.4 %, respectively (Faure and Mensing, 2005). Of these, only 204Pb is not radiogenic and its absolute 107 abundance has been stable since the formation of the Earth, while the other three (206Pb, 207Pb and 208Pb) 108 are formed as end products of radioactive decay from three nuclides, 238U, 235U and 232Th, respectively. The 109 three ratios of each radiogenic isotope to 204Pb significantly aid in determining the geochronological age of 110 Page 4 of 23 ore minerals from different deposits. Mercury (Hg) has seven stable isotopes (196−204 amu), which can be 111 affected by both mass-dependent fractionation (MDF, expressed as δ202Hg) and mass-independent 112 fractionation (MIF; expressed as ∆199Hg and ∆201Hg). The MDF of Hg is known to occur during 113 reduction−oxidation reactions, biological cycling, and volatilization of Hg (Blum and Bergquist, 2007). 114 When Hg isotopes undergo MIF, evenand odd-mass-numbered isotopes fractionate from each other; this 115 is most pronounced during UV-induced photochemical reactions (Bergquist and Blum, 2007). Recent 116 research explored the sources of pre-Inca and colonial archaeological cinnabar, as well as the Hg pollution 117 preserved in lake sediment cores in Peru and the Galápagos Islands (Cooke et al., 2013). This research 118 suggested that cinnabar ores from the largest cinnabar-bearing district in Central and Southern America 119 (Huancavelica, Peru) were used. In contrast, cinnabar for Inca artifacts originated from a distinctly different 120 source. Such evaluations can complement the provenance analysis of ores used to manufacture 121 archaeological artifacts. 122 The aim of this study was to analyze the origin and use of cinnabar in Roman wall paintings found in both 123 private and public houses at Ephesus of the mid-1st to the early 3rd century CE through mineralogical-124 petrographic and geochemical analyses. This aimed to shed light on changes in material supply and 125 processing, and with this, painting activities organization across different architectural units (likely in a 126 diachronic approach). The evaluation of the origin of the cinnabar used will be aided by making 127 comparisons to available reference data and newly acquired information on cinnabar deposits from the west 128 coast of Turkey, specifically from the Karaburun Peninsula. Overall, this work contributes to the discussion 129 of the organization of pigment trade and processing practices in antiquity. 130 131 132 2. MATERIALS AND METHODS 133 2.1. Geological materials 134 The Karaburun cinnabar mining district consists of three closely associated deposits: Kalecik, Dikencik and 135 Karareis (Fig. 1). These deposits are in the northern part of the Karaburun Peninsula, which is west of İzmir 136 and north of Ephesus. This study includes six samples from these deposits: four from Kalecik and two from 137 Dikencik (examples can be seen in Fig. 2). 138 Mercury deposits occur in a large variety of rock types, often associated with structural discontinuities. This 139 is also the case for the Karaburun deposits, which can be observed in various geological formations. The 140 oldest rock units in the northern part of the Karaburun Peninsula are represented by relatively autochthonous 141 successions (~495–292 Ma) that consist of clastic sedimentary rocks (Küçükbahçe and Dikendağı 142 Formations) at the western side and various tectonostratigraphic allochthonous units (~245–65 Ma) 143 consisting mainly of carbonates (Gerence, Camiboğazı, Güvercinlik Formations) and flysch deposits (İzmir 144 flysch and Yeniliman serpentinite) at the eastern side (Çakmakoğlu and Bilgin, 2006). These units are 145 unconformably covered by younger rock units of Neogene and Quaternary age (~24 Ma–present) and 146 include volcanic and sedimentary successions (Helvacı et al., 2009; Çakmakoğlu et al., 2013). The ore 147 deposits are epithermal and primarily contain mercury sulphide and quartz crystallized by rising 148 Page 5 of 23 hydrothermal solutions. The main ore mineral is cinnabar, which occurs as veinlets and fracture fillings in 149 silicified wall rocks (Yıldız and Bailey, 1978). The origin of the cinnabar occurrences is likely related to 150 Neogene volcanism (Yıldız et al., 1967; Helvacı et al., 2009), which introduced mercury-rich solutions into 151 the structural discontinuities of the various formations in the Karaburun Peninsula. 152 The Kalecik deposit is situated along a channelway through which the rising solutions flowed, following 153 the major northwest-trending fault cutting through the İzmir flysch rocks. The İzmir flysch rocks consist of 154 a sandstone-mudstone dominated matrix and radiolarites, basic volcanic rocks, ultrabasic rocks, and 155 limestone blocks. The ore was likely deposited just beneath the basalt flow unit (Yıldız and Bailey, 1978; 156 Helvacı et al., 2009). In the Dikencik deposit, cinnabar is observed in quartz veins along a fault that silicified 157 the Dikendağı Formation, which includes sand-, siltand mudstones with black cherts (lydite). The ore zone 158 is approximately 0.7 m wide and 75 m long and is unique in Turkey as it contains a considerable amount 159 of native mercury (Yıldız and Bailey, 1978). The Karareis deposit is very close to the Dikencik mine and 160 is present in the clastic sediments of the Dikendağı Formation. Due to the close proximity to the Dikencik 161 mine and due to cost considerations, no samples from the Karareis deposit were included in this study. The 162 geological materials were collected during fieldwork in 2013 (Fig. 1 and 2). Subsamples of these ore 163 samples were processed for isotopic analysis at Vrije Universiteit Brussel, Université Libre de Bruxelles, 164 and Ghent University, Belgium. 165 166 2.2. Archaeological materials 167 The samples drawn from upper painted layers of plasters (as well as cross-sections) were collected from 168 wall painting fragments stored in various boxes at the Ephesus Excavation House of the Austrian 169 Archaeological Institute of the Austrian Academy of Sciences, in Turkey. According to excavation notes, 170 these fragments belong to the southern portal of the Tetragonos Agora (referred to as Agora), as well as to 171 Terrace House 1 and Terrace House 2 (Fig. 3). Some excavation notes also attribute the fragments to 172 specific living units and rooms. However, precise contextualization and dating are challenging. This work 173 uses the excavation notes, descriptions of living units/rooms, and available information about building 174 periods, to provide all the available information about the samples. The samples were collected from wall 175 painting fragments from Terrace House 1 (Taberna IV), various living units and rooms of Terrace House 2 176 (including a taberna adjacent to Living Unit 7), and wall painting fragments from the Agora. However, the 177 samples were not collected from walls still in situ in various Roman houses introducing an uncertainty 178 regarding building structure and periods. These samples more generally represent cinnabar pigments from 179 Roman Ephesus. 180 The top layer of wall painting fragments (e.g., Fig. 2) was first analyzed using the handheld energy-181 dispersive X-ray fluorescence (HH-EDXRF) spectrometer Olympus InnovX Delta Premium 6000 (Rh 182 anode, 8-40 keV, Si-drift detector, 4W X-ray tube, current range of 5-200 μA). Based on this semi-183 quantitative in situ analysis of major and trace elements, wall painting fragments were selected on-site for 184 sample collection and further analysis. 185 The insula houses Terrace House 1 and Terrace House 2 were excavated amidst the remains of Ephesus. 186 Terrace House 1 covers an area of ~3,000 m2 and consists of several living units, each containing multiple 187 Page 6 of 23 rooms. Terrace House 2 covers an area of ~4,000 m2; it consists of seven living units. Both insula houses 188 were built in the first century CE and subsequently destroyed during an earthquake in 262/263 CE (Thür 189 and Rathmayr, 2014; Delile et al., 2015). The samples from Terrace House 1 pertain to Taberna IV. 190 Renovation works were undertaken in Terrace House 1 following an earthquake (2nd period, ~17–140 CE; 191 Lang-Auinger and Outschar, 1996). The samples are likely pertaining to this renovation period. The 192 samples from the Tetragonos Agora are from the Southern portal. The samples from Terrace House 2 193 pertain to the Taberna adjacent to Living Unit 7 (Room 45), as well as to Room 36 of Living Unit 6, Room 194 12 of Living Unit 3, and Room 21 of Living Unit 4. Wall paintings with cinnabar were discovered in Room 195 36 and 31b, leading to the proposal that these would represent an early painting phase (Thür and Rathmayr, 196 2014). The building phases in Terrace House 2 are categorized into four groups spanning from the mid-1st 197 c. CE to the early 3rd c. CE (Zimmermann, 2005). 198 Stratigraphic wall painting cross-sections, and scrapings of the top pigment layer, were collected with a 199 saw and scalpel, respectively. Of the wall painting fragments collected, ten (four from Terrace House 2, 200 two from Terrace House 1, and four from the Agora) were processed into polished petrographic cross-201 sections to study plaster and paint layer stratigraphy. All samples were obtained from large sections of 202 monochrome red wall paintings, although occasionally other hues were visible (e.g., EPHCIN-20 and -26; 203 Fig. 2). The pigment powders were processed at Vrije Universiteit Brussel, Université Libre de Bruxelles, 204 and Ghent University for isotopic analysis. 205 206 2.3. Cross-sections: optical microscopy and SEM-BSE/EDX 207 To study the morphology of cinnabar pigment and its application on painted surfaces, the plaster fragments 208 processed as cross-sections were examined using both light microscopy and SEM-BSE/EDX. This 209 examination was conducted on unprocessed fragments as well as on those prepared as stratigraphic polished 210 cross-sections. For incident light microscopy, a Nikon SMZ 1500 stereomicroscope was used. In addition 211 to optical microscopy, SEM analysis was performed under high vacuum conditions, using an accelerating 212 voltage of 15-25 kV in back-scattered electron (BSE) detection mode on the cross-sections to provide a 213 clearer view of cinnabar morphology and its distribution within the samples. Cinnabar appears brighter in 214 SEM-BSE images relative to red and/or yellow ochre owing to the higher atomic mass; this was exploited 215 to create images in which only the cinnabar grains are visible by radically adjusting the brightness/exposure 216 until the ochre grains no longer are visible. This can be verified since there are corresponding light 217 microscopy images showing the location of both cinnabar and ochre pigment grains concentrated in 218 separate paint layers in the same samples; once the known ochre layers disappear from the image, it is 219 assumed that only cinnabar remains in the BSE image. This allows the exclusive study of cinnabar, 220 including quantitative grain size and distribution image analysis with Image-J freeware (which was also 221 performed on SEM-BSE images of the surface of painting fragments before processing into cross sections). 222 Semi-quantitative EDX analysis of 20 grains of cinnabar per sample, thus isolated, was performed in 223 support of the more precise and specific chemical analyses described below. The SEM-BSE/EDX analysis 224 was performed at the University of Applied Arts Vienna using a Quanta FEG 250 (FEI, U.S.A.) scanning 225 electron microscope coupled to the Octane Elect Plus EDX detector (Ametek/EDAX, U.S.A.) and equipped 226 with Genesis EDX Quant software. 227 Page 7 of 23 228 2.4. Isotopic analysis 229 Subsamples (a few mg) of the collected top pigment layers (consisting of cinnabar and red clays, see 3.1) 230 and the geological samples from the Karaburun deposits were fully digested in aqua regia under a 100-231 class flow hood at the G-Time laboratories at Université Libre de Bruxelles. Lead was chromatographically 232 isolated from the digested geological and archaeological samples using the anion exchange BioradTM AG1-233 X8 resin (mesh size of 100-200 µm). Lead was separated from the matrix elements using a standard protocol 234 based on the use of HBr-HCl (Vanderstraeten et al., 2020). After the purified Pb eluent fractions were 235 obtained and evaporated to dryness, they were dissolved in 100 µl of concentrated HNO3, evaporated and 236 finally dissolved in 0.05% HNO3, for subsequent isotopic analysis using a Nu Plasma II (Nu II) high-237 resolution multi-collector inductively-coupled-plasma mass-spectrometry (HR-MC-ICP-MS) instrument 238 from Nu Instruments at the Laboratoire G-time, Université Libre de Bruxelles. 239 A thallium standard solution (Alpha ICP standard) used as a dopant - with a known 205Tl/203Tl isotope ratio 240 of 2.3871 - was added to all sample and standard solutions. Solutions were prepared to obtain a Pb-Tl ratio 241 of 4 or 5, and a minimum signal of 100 mV in the axial collector for 204Pb (i.e., the concentrations were 242 controlled to get 200 ppb of Pb and 50 ppb of Tl). A combination of internal correction based on the 243 monitoring of the Tl isotope ratio and external correction with the standard measured in a sample-standard 244 bracketing (SSB) approach was applied to correct for the bias induced by instrumental mass discrimination 245 for the measured Pb isotope ratios. Isobaric interference on mass 204 due to potentially present 204Hg was 246 monitored during acquisition by measuring the signal intensity at an m/z = 202 (Weis et al., 2006; 247 Vanderstraeten et al., 2020). While the Pb total beam was at ~8 V, the 202Hg beam intensity remained below 248 3 mV. The corrected Pb isotope ratios are reported using the recommended values of Abouchami et al. 249 (2000) for the Pb isotopic reference material NIST SRM 981. 250 A solution of the isotopic reference material NIST SRM 981 Pb standard solution was routinely analyzed 251 before starting a sample batch, as well as between every two samples to assess precision and reproducibility; 252 repeated analyses of NIST SRM 981 provided mean values of 16.941 ± 0.004, 15.501 ± 0.005 and 36.716 253 ± 0.015 (2SD, n=26) for 206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb, respectively. These results are in line with 254 the long-term repeatability of NIST SRM 981 values at the Laboratoire G-time of 16.9407 ± 0.0036, 255 15.4968 ± 0.0047 and 36.7163 ± 0.0121 (2SD, n=167; Vanderstraeten et al., 2020) for 206Pb/204Pb, 256 207Pb/204Pb and 208Pb/204Pb, respectively, and with the data reported by (Weis et al., 2005). Six samples were 257 analyzed in duplicate (EPHCIN-17, -19, -20, -27; EPHCINO-04 and -05), and the data repeatability was 258 within two standard deviations (2SD) for all samples. The procedural blank contained a Pb amount of 5 ng 259 (n=1), which is significantly lower (<5%) than the sample Pb contents and thus not expected to affect the 260 Pb isotope ratio results. The blank contribution to individual samples was <0.1%. 261 The isotopic analysis of Hg was carried out using a Neptune MC-ICP-MS unit from Thermo Fisher 262 Scientific at Ghent University. Hg was introduced as Hg(0) generated via the selective reduction of Hg2+ 263 with 3% SnCl2.2H2O in 1.2 M HCl in an HGX-200 cold vapor and hydride generation unit (Teledyne Cetac 264 Technologies, USA). The Hg(0)-loaded carrier gas coming from the cold-vapor generation (CVG) unit was 265 admixed in a ‘T’ piece with a wet aerosol of Tl generated by using a 100 µL min-1 concentric nebulizer 266 Page 8 of 23 mounted onto a dual (cyclonic and Scott-type) spray chamber (this setup is described in detail elsewhere: 267 Rua-Ibarz et al., 2016a,b, 2019; Bolea-Fernandez et al., 2019). 268 For instrumental mass discrimination correction, like for Pb isotopic analysis, a combination of internal 269 correction using Tl as admixed internal standard (NIST SRM 997 - Tl) in a “Baxter approach” and external 270 correction (NIST SRM 3133 - Hg) with the SSB-approach was used (Baxter et al., 2006). An in-house 271 standard solution of Hg previously characterized for its isotopic composition was measured in-between the 272 samples for Quality Assurance and Quality Control (QA/QC). Using NIST SRM 3233 (Hg), the external 273 precision was calculated to be ≤0.12 ‰ (2SD; over a period of 18 months, n=250; Rua-Ibarz et al., 2016a). 274 This precision is sufficient for studying the natural isotope ratio variation of Hg. To obtain accurate results, 275 the Hg concentration and acid content of all samples and standards were matched within ± 10 %. Blank 276 subtraction was not applied because the effect of the blank was demonstrated to be negligible within the 277 experimental precision. 278 Hg isotope ratios are reported as delta (δxxxHg) values. To separate the effect of MIF on the Hg isotope ratio 279 from that of MDF, the contribution caused by MDF is subtracted from the experimentally determined 280 δxxxHg values using the equations below. Both delta and capital delta values are reported in per mil (‰) 281 (Blum and Bergquist, 2007). 282 283 284 where xxx = 199, 200, 201 or 202 and NIST SRM 3133 is the Hg isotopic reference material. 285 286 287 288 289 290 3. RESULTS 291 3.1. Material processing, pigment layers and painting techniques 292 Visual examination of the wall painting fragments from Ephesus and optical microscopy (OM) shows a 293 similar appearance. However, a closer examination beneath the surface of the 10 cross-section samples 294 reveals technical details not immediately evident through visual inspection or non-invasive analysis. Figure 295 4 shows four examples of these cross-sections as examined by OM, along with SEM/EDX micrographs of 296 pigment layers with the exposure adjusted to highlight the cinnabar grains only. These images display four 297 different painting applications / cinnabar usage described below; each of the 10 cross sections fit into this 298 typology. They are: 299 Technique 1 (Agora) (samples EPHCIN-26, -28): Figure 4A shows a sample taken from the Tetragonos 300 Agora, where a thin (± 25 µm) layer of cinnabar was applied over an ochre layer applied a secco - ochre 301 Page 9 of 23 paint has seeped into drying cracks. The granulometry of this sample also indicates the use of a coarser 302 pigment (average particle size 8.5 µm, site average of 5.8 µm; Tab. 1). 303 Technique 2 (TH1) (samples EPHCIN-20, -24): Figure 4B, representing a sample from Terrace House 1, 304 displays a pigment layer over 100 µm thick, consisting of 2 layers of ochre, one yellow and one red, topped 305 by a thin layer of cinnabar. 306 Technique 3 (TH2A) (samples EPHCIN-05, -06, -14, -18): Figure 4C shows a sample from Terrace House 307 2, exhibits a clearly distinct technique from that of Terrace House 1; it shows a two-layer system of cinnabar 308 mixed with ochre over an ochre underpainting around 100 µm thick. 309 Technique 4 (TH2B (samples EPHCIN-21, -25): Figure 4D shows a second technique from Terrace House 310 2; it displays a thick (±75 µm) layer containing cinnabar pigment only, over an ochre-tinted intonaco layer. 311 Additionally, the sample in Figure 4D contains an ochre-tinted intonaco layer roughly 5 mm below the 312 finished surface. 313 Despite the difference in ochre underpainting between Techniques 2 and 4, there are also some similarities. 314 The visual resemblance is apparent in both techniques (Fig. 4B, D), which use a significant amount of 315 cinnabar, but digital image analysis reveals that the grain size distribution of the cinnabar is quite similar. 316 In contrast, Technique 3 (Fig. 4C) shows that not only was less cinnabar used, which is visually apparent 317 as well as supported by an average particle spatial distribution of 7 µm, but it also tended to be a finer grade 318 of pigment (Tab. 1). 319 320 3.2. Element and isotope ratio data 321 EDX analysis on a representative number of cinnabar grains in all 10 cross-sectioned samples, provided 322 semi-quantitative elemental characterization. The stoichiometry of the cinnabar used at the site appears 323 relatively similar, with any differences in the mercury-to-sulphur balance attributed to other elemental 324 “impurities” affecting this balance. Trace amounts of As and Se are present in the cinnabar grains of the 325 samples analyzed, at slightly different levels from sample to sample. Broadly speaking, in samples of 326 painting Technique 3 the signal for As and Se was at or below the margin of error for the detection limit, 327 i.e., they could not be determined to be present. Samples from Technique 4 had the highest levels of these 328 elements detected, equivalent to about a ratio of 200:1 S:As, 400:1 S:Se – too low for definite quantification, 329 but far enough above the background signal to be determined to be present. The cinnabar grains of the other 330 two techniques (1 and 2) had average levels of As and Se roughly half that of Technique 4. An important 331 caveat to this is that some grains had no As or Se detectable, while others had levels as high as Technique 332 4, which may indicate pigment mixing. However, due to the low levels detected and limitations of the 333 method, it is generally advisable to defer to the results of geochemical testing described that follows. 334 Despite the possibility for contamination in these samples and others, it is expected that the Pb and Hg 335 contents (and their respective isotope ratios) accurately reflect the visually identified cinnabar. 336 The Pb isotope ratios of the Karaburun ores (n=8) fall within the range 18.96 - 19.46 (± 0.01, 2SD) for 337 206Pb/204Pb, 15.70 - 15.72 (± 0.01, 2SD) for 207Pb/204Pb, and 38.62 - 38.99 (± 0.02, 2SD) for 208Pb/204Pb 338 (Tab. 2). The Pb isotope ratios of the Ephesian pigment samples (n=18) vary between 18.24 and 18.82 (± 339 0.01, 2SD) for 206Pb/204Pb, 15.63 and 15.69 (± 0.01, 2SD) for 207Pb/204Pb, and between 38.27 and 38.96 (± 340 Page 16 of 23 Data and materials availability: All data used for the study is included in this paper. Materials used in the 583 analysis are available upon request and depending on material availability. 584 585 REFERENCES 586 Abouchami, W., Galer, S.J.G., Hofmann, A.W., 2000. High precision lead isotope systematics of lavas 587 from the Hawaiian Scientific Drilling Project. Chemical Geology 169, 187–209. 588 https://doi.org/10.1016/S0009-2541(00)00328-4 589 Artioli, G., Angelini, I., Nimis, P., Villa, I.M., 2016. A lead-isotope database of copper ores from the 590 Southeastern Alps: a tool for the investigation of prehistoric copper metallurgy. Journal of 591 Archaeological Science 75, 27–39. https://doi.org/10.1016/j.jas.2016.09.005 592 Baragona, A.J., Rodler, A.S., Bauerová, P., 2022. Imperial Styles, Frontier Solutions: Roman wall 593 painting technology in the province of Noricum. In: Proceedings of the 6th Historic Mortars 594 Conference, HMC2022 University of Ljubljana, SL. RILEM Sarl, France 595 Baxter, D.C., Rodushkin, I., Engström, E., Malinovsky, D., 2006. Revised exponential model for mass 596 bias correction using an internal standard for isotope abundance ratio measurements by multi-597 collector inductively coupled plasma mass spectrometry. Journal of Analytical Atomic 598 Spectrometry, 21, 427–430. https://doi.org/10.1039/B517457K 599 Beck, L.Y., 2005. De Materia Medica. Olms-Weidmann: Hildesheim. 600 Becker, H. 2021. Color Technology and Trade. In: Biggam, C. and Wolf, K. and Wharton, D. (eds) 601 Cultural History of Color, ed. Bloomsbury Publishing, London, 35-48. 602 Becker, H. 2022. Pigment nomenclature in the ancient Near East, Greece, and Rome. Journal of 603 Archaeological and Anthropological Sciences 14:20. DOI: 10.1007/s12520-021-01394-1 604 Becker, H. Commerce in color: the mechanics of the Roman pigment trade. forthcoming 605 Begemann, F., Schmitt-Strecker, S., Pernicka, E., 2003. On the Composition and Provenance of Metal 606 Finds from Beşiktepe (Troia). In: Wagner, G.A., Pernicka, E., Uerpmann, HP. (eds) Troia and the 607 Troad. Natural Science in Archaeology. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-608 3-662-05308-9_11 609 Bergquist, B. A., Blum, J. D., 2007. Mass-dependent and-independent fractionation of Hg isotopes by 610 photoreduction in aquatic systems. Science 318 (5849), 417. 611 Blum, J.D., Bergquist, B.A., 2007. Reporting of variations in the natural isotopic composition of mercury. 612 Analytical and Bioanalytical Chemistry 388, 353–359. https://doi.org/10.1007/s00216-007-1236-9 613 Page 17 of 23 Bolea-Fernandez, E., Rua-Ibarz, A., Krupp, E.M., Feldmann, J., Vanhaecke, F., 2019. High-precision 614 isotopic analysis sheds new light on mercury metabolism in long-finned pilot whales (Globicephala 615 melas). Sci Rep 9, 7262 (2019). https://doi.org/10.1038/s41598-019-43825-z 616 Bracci, S. Cantisiani, E., Falzone, F., Marano, M., Tomassini, P., 2021. Archaeometry and Roman Wall 617 Painting: The Case of pre-Hadrianic Paintings in Ostia Antica. In: Cavalieri and Tomassini (ed.) Le 618 peinture murale antique: méthodes et apports d’une approche technique Actes du colloque 619 international. Louvain la Neuve 21 April 2017 Edizioni Quasar 620 Çakmakoğlu, A., Bilgin, Z.R., 2006. Pre-Neogene Stratigraphy of The Karaburun Peninsula (W of İzmir 621 Turkey). Bulletin of The Mineral Research and Exploration 132, 33–61. 622 Çakmakoğlu, B., Göktaş, F., Demirhan, M., Helvacı, C., 2013. Investigation of Stratigraphy, 623 Sedimentology and Economical Usage of the Clays from the Northern Part of the Karaburun 624 Peninsula. Geological Bulletin of Turkey 56(1), 39–58. 625 Chen, D., Ren, D., Deng, C., Tian, Z., Yin, R., 2022. Mercury loss and isotope fractionation during high-626 pressure and high-temperature processing of sediments: Implications for the behaviors of mercury 627 during metamorphism. Geochimica et Cosmochimica Acta 334, 231–240. 628 https://doi.org/10.1016/j.gca.2022.08.010 629 Cooke, C.A., Hintelmann, H., Ague, J.J., Burger, R., Biester, H., Sachs, J.P., Engstrom, D.R., 2013. Use 630 and legacy of mercury in the Andes. Environmental Science & Technology 47, 4181−4188. 631 https://doi.org/10.1021/es3048027 632 Delile, H., Blichert-Toft, J., Goiran, J.P., Stock, F., Arnaud-Godet, F., Bravard, J.P., Brückner, H., 633 Albarède, F., 2015. Demise of a harbor: A geochemical chronicle from Ephesus. Journal of 634 Archaeological Science 53, 202-213. https://doi.org/10.1016/j.jas.2014.10.002 635 Emslie, S.D., Brasso, R., Patterson, W.P., Valera, A.C., McKenzie, A., Silva, A.M., Gleason, J.D., Blum, 636 J.D., 2015. Chronic mercury exposure in Late Neolithic/Chalcolithic populations in Portugal from 637 the cultural use of cinnabar. Scientific Reports 5, 14679. DOI: 10.1038/srep14679 638 Esposito, D., 2014. La pittura di Ercolano. Studi della Soprintendenza Archeologica di Pompeii 33. 639 Rome: “L’Erma” di Bretschneider 640 Esposito, D., 2017. The Economics of Pompeian Painting. In: Flohr, M. and Wilson, A. (eds.), The 641 Economy of Pompeii (pp. 263-289). Oxford: Oxford University Press. 642 Faure, G., Mensing, T.M., 2005. Isotopes, Principles and Applications. 3rd ed. Hoboken Wiley. 643 Foss, C., Hanfmann, G.M.A., 1975. Reginal setting and urban development. In: Hanfmann, G.M.A., 644 Jacobs, S.W. (eds.), Achraological Exploraiton of Sardis: A survey of Sardis and the major 645 monuments outside the city walls. Harvard University Press, 17–34. 646 Page 18 of 23 Fragnoli, P., Ugarković, M., Sterba, J.H., Sauer, R., 2022. Looking for Ephesian workshops: an integrated 647 petrographic, geochemical, and chrono-typological approach to Late Hellenistic Ephesos lamps. 648 Archaeological and Anthropological Sciences 14, 19. https://doi.org/10.1007/s12520-021-01419-9 649 Gale, N.H., Stos-Gale, Z.A., 1981. Lead and silver in the ancient Aegean. Scientific American 244(6), 650 176–192. 651 Gemici, Ü., 2008. Batı Anadolu’daki eski civa yataklarının çevre jeolojisi açısından değerlendirilmesi. In: 652 Eşref Atabey (ed.) Uluslararası Katılımlı Tıbbi Jeoloji Sempozyumu Kitabı. 133–145. 653 Gemici, Ü., Oyman, T., 2003. The influence of the abandoned Kalecik Hg mine on water and stream 654 sediments (Karaburun, Izmir, Turkey). Science of the Total Environment 312, 155–166. 655 doi:10.1016/S0048-9697(03)00008-1 656 Gliozzo, E., 2021. Pigments — Mercury-based red (cinnabar-vermilion) and white (calomel) and their 657 degradation products. Archaeological and Anthropological Sciences 13, 210. 658 https://doi.org/10.1007/s12520-021-01402-4 659 Gökce, A., Bozkaya, G., 2007. Lead and sulfur isotopic studies of the barite–galena deposits in the 660 Karalar area (Gazipaoa–Antalya), Southern Turkey. Journal of Asian Earth Sciences 30, 53–62. 661 doi:10.1016/j.jseaes.2006.07.007 662 Gray, J.E., Pribil, M.J., Higueras, P.L., 2013. Mercury isotope fractionation during ore retorting in the 663 Almadén mining district, Spain. Chemical Geology 357, 150–157. 664 https://doi.org/10.1016/j.chemgeo.2013.08.036 665 Hauptmann, A., Schmitt-Strecker, S., Begemann, F., Palmieri, A.M., 2002. Chemical Composition and 666 Lead Isotopy of Metal Objects from the "Royal" Tomb and Other Related Finds at Arslantepe, 667 Eastern Anatolia. Paléorient 28, 43–69. http://www.jstor.org/stable/41496946 668 Helvacı, C. Ersoy, E.Y., Sözbilir, H., Erkül, F., Sümer, Ö., Uzel, B., 2009. Geochemistry and 40Ar/39Ar 669 geochronology of Miocene volcanic rocks from the Karaburun Peninsula: Implications for 670 amphibole-bearing lithospheric mantle source, Western Anatolia. Journal of Volcanology and 671 Geothermal Research 185. 181–202. 672 Higueras, P., Munhá, J., Oyarzun, R., Tassinari, C.C.G., Ruiz, I.R., 2005. First lead isotopic data for 673 cinnabar in the Almadén district (Spain): implications for the genesis of the mercury deposits. 674 Mineralium Deposita 40, 115-122. https://doi.org/10.1007/s00126-005-0471-2 675 Hirao, Y., Enomoto, J., Tachikawa, H., 1995. Lead isotope ratios of copper, zinc and lead minerals in 676 Turkey – in relation to the provenance study of artifacts. In: Prince Takahito Mikasa HIH (ed.) 677 Essays on Ancient Anatolia and its surrounding civilisations. Harrassovitz Verlag, Wiesbaden, 89–678 114. 679 Hunt-Ortiz, M.A., Consuegra-Rodríguez, S., Díaz del Río-Español, P., Hurtado-Pérez, V.M., Montero-680 Ruiz, I., 2011. Neolithic and Chalcolithic – VI to III Millennia BC – use of cinnabar (HgS) in the 681 Page 19 of 23 Iberian Peninsula: analytical identification and lead isotopic analysis for an early mineral 682 exploitation of the Almadén (Ciudad Real, Spain) mining district. In: J. E. Ortiz, O. Puche, I. 683 Rábano and L. F. Mazadiego (eds.) History of Research in Mineral Resources. Cuadernos del 684 Museo Geominero, 13. Instituto Geológico y Minero de España, Madrid. 685 Keil, J., von Premerstein, A., 1914. Bericht über eine 3. Reise in Lydien und den angrenzenden Gebieten 686 Ioniens, ausgeführt 1911 im Auftrage der Kaiserlichen Akademie der Wissenschaften (Wien 1914). 687 Klein, S., Domergue, C., Lahaye, Y., Brey, G.P., von Kaenel, H.-M., 2009. The lead and copper isotopic 688 composition of copper ores from the Sierra Morena (Spain). J. Iber. Geol. 35, 59–68. 689 Klein, S., Lahaye, Y., Brey, G.P., von Kaenel, H.-M., 2004. The early Roman Imperial aes coinage II: 690 tracing the copper sources by analysis of lead and copper isotopes – copper coins of Augustus and 691 Tiberius. Archaeometry 46(3), 469–480. 692 Kostomitsopoulou Marketou, A., 2019. The pigment production site of the ancient agora of Kos (Greece): 693 Revisiting the material evidence. THIASOS: Rivista di archeologia e architettura antica 8, 61–80. 694 Kostomitsopoulou Marketou, A., Andriulo, F., Steindal, C., Handberg, S., 2020. Egyptian blue pellets 695 from the first century B.C.E workshop of Kos (Greece): Microanalytical investigation by Optical 696 Microscopy, Scanning Electron Microscopy-X-ray Energy Dispersive Spectroscopy and Micro-697 Raman Spectroscopy. Minerals 10, 1063. https://doi.org/10.3390/min10121063 698 Kostomitsopoulou Marketou, A., Rodler-Rørbo, A., 2023. The Colourant Mapping Project: approaching 699 the evidence of colourant production in the Aegean during the Hellenistic and Roman Period 700 through Kea, Kos and Rhodes. Archaeological and Anthropological Sciences 15, 160. 701 https://doi.org/10.1007/s12520-023-01849-7 702 Ladstätter, S., 2016. Hafen und Stadt von Ephesos in hellenistischer Zeit. Jahreshefte des 703 Österreichischen Archäologischen Institutes in Wien 85, 233–272. ISSN 2309-1207 704 Lang-Auinger, C., 1996. Hanghaus 1 in Ephesos. Der Baubefund. Jahreshefte des Österreichischen 705 Archäologischen Institutes in Wien VIII/3. 706 Lang-Auinger, C., Outschar, U., 1996. Aspekte zur Chronologie. In: Lang-Auinger, C. (ed.), Hanghaus 1 707 in Ephesos. Der Baubefund. Jahreshefte des Österreichischen Archäologischen Institutes in Wien 708 VIII/3. 709 Lazzarini, L., Verità, M., 2015. First evidence for 1st century AD production of Egyptian blue frit in 710 Roman Italy. Journal of Archaeological Science 53, 578–585. 711 Maras, A., Botticelli, M., Ballirano, P., 2013. Archaeometric investigations on cinnabar provenance and 712 origin by X-ray powder diffraction: preliminary data. International Journal of Conservation Science 713 4, 685–692. 714 Page 20 of 23 Mazzocchin, G. A., Baraldi, P., Barbante, C., 2008. Isotopic analysis of lead present in the cinnabar of 715 Roman wall paintings from the Xth Regio “(Venetia et Histria)” by ICP‐MS. Talanta 74, 690–693. 716 https:// doi.org/10.1016/j.talanta.2007.06.048 717 Minami, T., Takeuchi, A., Imazu, S., Okuyama, M., Higashikage, Y., Mizuno, T., Okabayashi, K., 718 Takahashi, K., 2021. Identification of source mine using sulfur, mercury, and lead isotope analyses 719 of vermilion used in three representative tombs from Kofun period in Japan. Journal of 720 Archaeological Science: Reports 37, 102970. https://doi.org/10.1016/j.jasrep.2021.102970 721 Ni, X., Yang, R., Yuan, W., Wang, X., Chen, J., Zhang, G., Li, D., Du, L., Gao, L., Luo, C., Zheng, L., 722 Xu, H., 2022. New insight into the source of metals in Hg deposits at the southwestern margin of 723 the Yangtze Platform, China: Evidence from mercury stable isotope compositions. Ore Geology 724 Reviews 149, 105089. https://doi.org/10.1016/j.oregeorev.2022.105089 725 Nicholson, P.T., 2003. New excavations at a Ptolemaic-Roman faience factory at Memphis, Egypt. 726 Annales du 15e Congres de l’Association Internationale pour l’Histoire du Verre. AIHV 49–52. 727 Palero-Fernández, F.J., Martin-Izard, A., Prieto, M.Z., Mansilla-Plaza, L., 2015. Geological context and 728 plumbotectonic evolution of the giant Almadén Mercury Deposit. Ore Geology Reviews 64, 71–88. 729 https://doi.org/10.1016/j.oregeorev.2014.06.013 730 Pribil, M.J., Rimondi, V., Costagliola, P., Lattanzi, P., Rutherford, D.L., 2020. Assessing mercury 731 distribution using isotopic fractionation of mercury processes and sources adjacent and downstream 732 of a legacy mine district in Tuscany, Italy. Applied Geochemistry 117, 104600. 733 https://doi.org/10.1016/j.apgeochem.2020.104600 734 Prieto, G., Wright, V., Burger, R.L., Cooke, C.A., Zeballos-Velasquez, E.L., Watanave, A., Suchomel, 735 M.R., Suescun, L., 2015. The source, processing and use of red pigment based on hematite and 736 cinnabar at Gramalote, an early Initial Period (1500–1200 cal. B.C.) maritime community, north 737 coast of Peru. Journal of Archaeological Science: Reports 5, 45–60. 738 http://dx.doi.org/10.1016/j.jasrep.2015.10.026 739 Rodler, A. S., Artioli, G., Klein, S., Petschick, R., Fink‐Jensen, P., Brøns, C., 2017. Provenancing ancient 740 pigments: Lead isotope analyses of the copper compound of Egyptian blue pigments from ancient 741 Mediterranean artifacts. Journal of Archaeological Science: Reports 16, 1–18. https://doi. 742 org/10.1016/j.jasrep.2017.09.008 743 Rodríguez, J, Montero‐Ruiz, I, Hunt‐Ortiz, M, García‐Pavón, E., 2020. Cinnabar provenance of 744 Chalcolithic red pigments in the Iberian Peninsula: A lead isotope study. Geoarchaeology, 1–12. 745 https://doi.org/10.1002/gea.21810 746 Rua-Ibarz, A., Bolea-Fernandez, E., Maage, A., Frantzen, S., Sanden, M., Vanhaecke, F., 2019. Tracing 747 mercury pollution along the Norwegian Coast via elemental, speciation, and isotopic analysis of 748 liver and muscle tissue of deep-water marine fish (Brosme brosme). Environmental Science & 749 Technology 53, 1776–1785. https://doi.org/10.1021/acs.est.8b04706 750 Page 21 of 23 Rua-Ibarz, A., Bolea-Fernandez, E., Maage, A., Frantzen, S., Valdersnes, S., Vanhaecke, F., 2016b. 751 Assessment of Hg pollution released from a WWII submarine wreck (U-864) by Hg isotopic 752 analysis of sediments and Cancer pagurus tissues. Environmental Science & Technology 50, 753 10361–10369. DOI: 10.1021/acs.est.6b02128 754 Rua-Ibarz, A., Bolea-Fernandez, E., Vanhaecke, F., 2016a. An in-depth evaluation of accuracy and 755 precision in Hg isotopic analysis via pneumatic nebulization and cold vapor generation multi-756 collector ICP-mass spectrometry. Analytical and Bioanalytical Chemistry, 408, 417−429. 757 DOI:10.1007/s00216-015-9131-2 758 Salvadori, M., Sbrolli, C., 2021. Wall paintings through the ages: the roman period – Republic and early 759 Empire. Archaeological and Anthropological Sciences 13, 187. https://doi.org/10.1007/s12520-760 021-01411-3 761 Sayre, E.V., Joel, E.C., Blackman, M.J., Yener, K.A., Özbal, H., 2001. Stable lead isotope studies of 762 Black Sea Anatolian ore sources and related Bronze age and Phrygian artefacts from nearby 763 archaeological sites. Appendix: new Central Taurus ore data. Archaeometry 43, 77–115. 764 Scherrer, P., Trinkl, E., 2006. Die Tetragonos Agora in Ephesos. Grabungsergebnisse von archaischer bis 765 in byzantinische Zeit – Ein Überblick Befunde und Funde klassischer Zeit. Jahreshefte des 766 Österreichischen Archäologischen Institutes in Wien XIII/2. 767 Schmeisser, C., 1906. Bodenschätze und Bergbau Kleinasiens. Zeitschrift für praktische Geologie XIV, 6, 768 186-196. 769 Smith, R.S., Wiederhold, J.G., Jew, A.D., Brown Jr., G.E., Bourdon, B., Kretzschmar, R., 2014. Small-770 scale studies of roasted ore waste reveal extreme ranges of stable mercury isotope signatures. 771 Geochimica et Cosmochimica Acta 137, 1–17. https://doi.org/10.1016/j.gca.2014.03.037 772 Spangenberg, J.E., Lavrič, J.V., Meisser, N., Serneels, V., 2010. Sulfur isotope analysis of cinnabar from 773 Roman wall paintings by elemental analysis/isotope ratio mass spectrometry – tracking the origin 774 of archaeological red pigments and their authenticity. Rapid communications in Mass Spectrometry 775 24(19), 2812-2816. https://doi.org/10.1002/rcm.4705 776 Stephens, J.A., Ducea, M.N., Killick, D.J., Ruiz, J., 2021. Use of non-traditional heavy stable isotopes in 777 archaeological research. Journal of Archaeological Science 127, 105334. 778 https://doi.org/10.1016/j.jas.2021.105334 779 Stetson, S.J., Gray, J.E., Wanty, R.B., Macalady, D.L., 2009. Isotopic Variability of Mercury in Ore, 780 Mine-Waste Calcine, and Leachates of Mine-Waste Calcine from Areas Mined for Mercury. 781 Environmental Science and Technology 43, 7331–7336. https://doi.org/10.1021/es9006993 782 Stos-Gale, Z.A., Gale, N.H., 2009. Metal provenancing using isotopes and the Oxford archaeological lead 783 isotope database (OXALID). Archaeological and Anthropological Sciences 1, 195–213. 784 https://doi.org/10.1007/s12520-009-0011-6 785 Page 22 of 23 Tezgör, D.K., 2022. From the miltos / sinopis of ancient Sinope to the Yoşa of Modern Cappadocia. 786 Adalya 25, 45–69. http://orcid.org/0000-0002-5736-1822 787 Thür, H., 2005. Hanghaus 2 in Ephesos – die Wohneinheit 4. Jahreshefte des Österreichischen 788 Archäologischen Institutes in Wien VIII/6. 789 Thür, H., Rathmayr, E., 2014. Hanghaus 2 in Ephesos – die Wohneinheit 6. Jahreshefte des 790 Österreichischen Archäologischen Institutes in Wien VIII/9. 791 Tober, B., 2021. Ephesos – Palmyra – Noricum. Room-decoration as an international Code for living with 792 images under the Roman imperial period. In: Thomas, R. (ed.), Local Styles or Common Pattern 793 Books in Roman Wall Painting and Mosaics (Panel 3.22), Proceedings of the 19th International 794 Congress of Classical Archaeology Cologne/Bonn 22 65-80 Heidelberg 22-26 May 2018 795 Tsantini, E., Minami, T., Takahashi, K., Ontiveros, M.Á.C., 2018. Analysis of sulphur isotopes to identify 796 the origin of cinnabar in the Roman wall paintings from Badalona (Spain). Journal of 797 Archaeological Science: Reports 18, 300-307. https://doi.org/10.1016/j.jasrep.2018.01.032 798 Vanderstraeten, A., Bonneville, S., Gili, S., de Jong, J., Debouge, W., Claeys, P., Mattielli, N., 2020. First 799 multi-isotopic (Pb-Nd-Sr-Zn-Cu-Fe) characterisation of dust reference materials (ATD and BCR-800 723): A multi-column chromatographic method optimised to trace mineral and anthropogenic dust 801 sources. Geostandards and Geoanalytical Research 44, 307–329. doi:10.1111/ggr.12320 802 Wagner, G.A., Pernicka, E., Seeliger, T.C., Öztunali, Ö., Baranyi, I., Begemann, F., Schmitt-Strecker, S., 803 1985. Geologische Untersuchungen zur frühen Metallugie in NW-Anatolien. Bulletin of the 804 Mineral Research and Exploration Inst. Turkey 101/102, 45–81. 805 Weis, D., Kieffer, B., Maerschalk, C., Barling, J., de Jong, J., Williams, G., Hanano, D., Pretorius, W., 806 Mattielli, N., Scoates, J., Goolaerts, A., Friedman, R., Mahoney, J., 2006. High-precision isotopic 807 characterization of USGS reference materials by TIMS and MC-ICP-MS: Isotopic study of USGS 808 reference materials. Geochemistry, Geophysics, Geosystems, 7(8), Q08006, 809 doi:10.1029/2006GC001283. 810 Weis, D., Kieffer, B., Maerschalk, C., Pretorius, W., Barling, J., 2005. High-precision Pb-Sr-Nd-Hf 811 isotopic characterization of USGS BHVO-1 and BHVO-2 reference materials. Geochemistry, 812 Geophysics, Geosystems, 6, Q02002, doi:10.1029/2004GC000852. 813 Willett, F., Sayre, E.V., 2006. Lead isotope in West African copper alloys. Journal of African Archaeology 814 4, 55–90. 815 Yener, K.A., Sayre, E.V., Joel, E.C., Özbal, H., Barnes, I.L., Brill, R.H., 1991. Stable lead isotope studies 816 of Central Taurus ore sources and related artifacts from Eastern Mediterranean Chalcolithic and 817 Bronze age sites. Journal of Archaeological Science 18, 541–577. 818 Page 23 of 23 Yesares, L., Sáez, R., De Almodóvar, G.R., Nieto, J.M., Gómez, G., Ovejero, G., 2017. Mineralogical 819 evolution of the Las Cruces gossan cap (Iberian Pyrite Belt): From subaerial to underground 820 conditions. Ore Geology Reviews 80, 377–405. https://doi.org/10.1016/j.oregeorev.2016.05.018 821 Yin, R., Feng, X., Hurley, J.P., Krabbenhoft, D.P., Lepak, R.F., Hu, R., Zhang, Q., Li, Z., and Bi, X., 822 2016. Mercury Isotopes as Proxies to Identify Sources and Environmental Impacts of Mercury in 823 Sphalerites. Scientific Reports 6, 18686. https://doi.org/10.1038/srep18686 824 Yin, R., Feng, X., Wang, J., Li, P., Liu, J., Zhang, Y., Chen, J., Zhen, L., Hu, T., 2013. Mercury 825 speciation and mercury isotope fractionation during ore roasting process and their implication to 826 source identification of downstream sediment in the Wanshan mercury mining area, SW China. 827 Chemical Geology 336, 72–79. https://doi.org/10.1016/j.chemgeo.2012.04.030 828 Yıldız, M., 1967. Mercury: Maden Tetkik ve Arama Enst. (MTA) Bull, 68. 829 Yıldız, M., Bailey, E.H., 1978. Mercury deposits in Turkey. Geological Survey Bulletin 1456, 1–79. 830 Zabrana, L., 2022. Das Artemision von Ephesos als Wirtschaftszentrum in römischer Zeit. Neue 831 archäologische Forschungen des Österreichischen Archäologischen Institutes. In: Alkier, S. (ed) 832 Konstellationen antiker Tempelwirtschaft, Beyond Historicism – New Testament Studies Today 833 Vol. 1. 834 Zimmermann, G.A., 1874. Ephesos im ersten christlichen Jahrhundert (Jena 1874). 835 Zimmermann, N., 2005. Wandmalerei. In: Thür, H. (ed.), Hanghaus 2 in Ephesos – die Wohneinheit 4. 836 Jahreshefte des Österreichischen Archäologischen Institutes in Wien VIII/6. 837 Zimmermann, N., 2014. Wandmalerei. In: Thür, H., Rathmayr, E. (eds.), Hanghaus 2 in Ephesos – die 838 Wohneinheit 6. Jahreshefte des Österreichischen Archäologischen Institutes in Wien VIII/9. 839 Zimmermann, N., 2016a. Wandmalerei. In: Rathmayr, E. (ed.), Hanghaus 2 in Ephesos – die Wohneinheit 840 7. Jahreshefte des Österreichischen Archäologischen Institutes in Wien VIII/10. 841 Zimmermann, N., 2016b. Die Entwicklung der Wandmalerei in Kleinasien von der griechischen bis zur 842 römischen Zeit am Beispiel von Ephesos. In: Dubois, Y., Niffeler, U. (eds.), Pictores per provincias 843 II – Status Quaestionis. Actes du 13e Colloque de l’Association Internationale pour la Peinture 844 Murale Antique (AIPMA), Université de Lausanne, 12–16 septembre 2016. 845 Table 1. Cinnabar grain size diameter, average particle size and intergranular spacing for the Ephesian wall painting fragments analyzed 1 Cinnabar grain size diameter Average particle size (µm) Average intergranular space (µm) Painting technique Sample ID Site <1 µm 1-2 µm 2-3 µm 3-5 µm 5-10 µm >10 µm (%) (%) (%) (%) (%) (%) 1–Agora EPHCIN-28 Agora n.d. n.d. 1.00 23.0 56.0 20.0 8.25 4.50 2–TH1 EPHCIN-24 Terrace House 1 3.00 19.0 14.0 24.0 32.0 8.00 4.70 3.10 3–TH2A EPHCIN-06 Terrace House 2, Room 36 n.d. 43.0 15.0 20.0 31.0 6.00 3.50 7.00 4–TH2B EPHCIN-25 Room 45, Taberna adjacent to living unit 7 of Terrace House 2 3.00 9.00 12.0 18.0 44.0 15.0 6.70 5.00 Average across samples 1.50 17.8 10.5 21.3 40.8 12.3 5.79 4.90 Notes: Tabular display of the information given in Fig. 1, with the additional information of the average cinnabar intergranular space; this is based on the 2 analysis of all cross-sectioned samples even though only the four shown in Fig. 1 are mentioned here as examples. 3 Table 2. Isotope ratio data for cinnabar samples from Ephesian wall painting fragments and cinnabar ore from the Karaburun Peninsula, Turkey 4 Sample ID 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb 207Pb/206Pb 208Pb/206Pb Δ199Hg δ 199 Hg Δ200Hg δ 200 Hg Δ201Hg δ 201 Hg δ 202 Hg EPHCIN-021 18.711 ± 0.001 15.684 ± 0.001 38.880 ± 0.002 0.83822 ± 0.00001 2.07791 ± 0.00004 Cinnabar, Ephesian wall painting fragments EPHCIN-061 18.771 ± 0.001 15.683 ± 0.001 38.819 ± 0.002 0.83549 ± 0.00002 2.06797 ± 0.00004 -0.13 ± 0.04 0.15 ± 0.05 -0.03 ± 0.06 0.54 ± 0.06 -0.10 ± 0.05 0.75 ± 0.05 1.13 ± 0.05 EPHCIN-131 18.686 ± 0.001 15.671 ± 0.001 38.818 ± 0.002 0.83865 ± 0.00001 2.07744 ± 0.00003 -0.09 ± 0.06 0.07 ± 0.08 0.01 ± 0.07 0.32 ± 0.11 -0.13 ± 0.06 0.34 ± 0.10 0.62 ± 0.10 EPHCIN-141 18.688 ± 0.001 15.671 ± 0.001 38.817 ± 0.002 0.83853 ± 0.00001 2.07707 ± 0.00004 EPHCIN-172* 18.818 ± 0.001 15.684 ± 0.001 38.864 ± 0.002 0.83348 ± 0.00002 2.06529 ± 0.00004 EPHCIN-191* 18.624 ± 0.001 15.672 ± 0.001 38.824 ± 0.002 0.84150 ± 0.00002 2.08457 ± 0.00004 EPHCIN-201* 18.613 ± 0.001 15.670 ± 0.001 38.805 ± 0.002 0.84183 ± 0.00002 2.08479 ± 0.00005 -0.13 ± 0.04 0.07 ± 0.03 0.05 ± 0.05 0.45 ± 0.05 -0.16 ± 0.07 0.43 ± 0.04 0.79 ± 0.08 EPHCIN-213 18.756 ± 0.001 15.681 ± 0.001 38.933 ± 0.003 0.83603 ± 0.00001 2.07582 ± 0.00005 EPHCIN-234 18.385 ± 0.001 15.648 ± 0.001 38.474 ± 0.002 0.85113 ± 0.00001 2.09270 ± 0.00004 EPHCIN-244 18.245 ± 0.001 15.635 ± 0.001 38.266 ± 0.002 0.85695 ± 0.00003 2.09733 ± 0.00006 -0.11 ± 0.05 0.08 ± 0.05 -0.02 ± 0.04 0.36 ± 0.04 -0.11 ± 0.07 0.46 ± 0.08 0.76 ± 0.05 EPHCIN-255 18.526 ± 0.001 15.656 ± 0.001 38.559 ± 0.002 0.84507 ± 0.00001 2.08135 ± 0.00004 -0.40 ± 0.06 -0.13 ± 0.07 -0.02 ± 0.04 0.52 ± 0.06 -0.33 ± 0.09 0.48 ± 0.09 1.07 ± 0.07 EPHCIN-266 18.784 ± 0.001 15.685± 0.001 38.958 ± 0.002 0.83504 ± 0.00002 2.07408 ± 0.00003 -0.31 ± 0.03 -0.15 ± 0.04 -0.04 ± 0.05 0.29 ± 0.04 -0.21 ± 0.08 0.29 ± 0.04 0.66 ± 0.08 EPHCIN-276* 18.768 ± 0.001 15.683 ± 0.001 38.942 ± 0.002 0.83563 ± 0.00001 2.07490 ± 0.00004 EPHCIN-286 18.760 ± 0.001 15.682 ± 0.001 38.907 ± 0.002 0.83591 ± 0.00001 2.07389 ± 0.00004 Karaburun cinnabar ore EPHCINO-03 D 19.034 ± 0.001 15.704 ± 0.001 38.986 ± 0.002 0.82501 ± 0.00001 2.04821 ± 0.00003 -0.05 ± 0.03 -0.04 ± 0.04 0.01 ± 0.05 0.04 ± 0.05 -0.07 ± 0.03 -0.03 ± 0.06 0.05 ± 0.06 EPHCINO-04D* 18.963 ± 0.001 15.701 ± 0.001 38.953 ± 0.002 0.82800 ± 0.00001 2.05414 ± 0.00004 EPHCINO-05K* 19.202 ± 0.001 15.706 ± 0.001 38.623 ± 0.002 0.81800 ± 0.00001 2.01149 ± 0.00004 EPHCINO-08 K 19.464 ± 0.001 15.721 ± 0.001 38.735 ± 0.002 0.80772 ± 0.00001 1.99012 ± 0.00004 -0.15 ± 0.11 -0.20 ± 0.13 0.00 ± 0.07 -0.12 ± 0.10 -0.11 ± 0.05 -0.29 ± 0.08 -0.23 ± 0.08 EPHCINO-11 K 19.332 ± 0.001 15.721 ± 0.001 38.746 ± 0.002 0.81320 ± 0.00001 2.00423 ± 0.00004 -0.22 ± 0.04 -0.19 ± 0.03 -0.02 ± 0.05 0.04 ± 0.06 0.02 ± 0.12 0.11 ± 0.15 0.12 ± 0.06 EPHCINO-12K 19.130 ± 0.001 15.710 ± 0.001 38.692 ± 0.002 0.82125 ± 0.00001 2.02261 ± 0.00003 BHVO-2 BHVO-2** 18.453 ± 0.001 18.62 ± 0.05 15.669 ± 0.001 15.53 ± 0.05 38.626 ± 0.003 38.21 ± 0.04 0.84915 ± 0.00002 2.09321 ± 0.00005 -0.15 ± 0.06 0.05 ± 0.06 0.03 ± 0.03 0.42 ± 0.04 -0.21 ± 0.06 0.38 ± 0.06 0.79 ± 0.09 Notes: Pb and Hg isotope ratios (±2SE) for samples of Terrace House 2: 1 Living unit 6 (Room 36a), 2 Living unit 3 (Room 12), 3 Living unit 4 (Room 21), 5 5 Room 45 (Taberna), Terrace House 1: 4 Taberna IV, and the 6 Agora; Karaburun ores EPHCINO-03D and -04D, and EPHCINO-05K, -08K, -11K and -12K are 6 from the Dikencik and Kalecik deposits, resp. (Fig. 1); *mean values (n=2); ** Pb isotope ratio data for reference material BHVO-2 (±2SD; Weis et al., 2005)7 Page 9 of 10 49 Figure 6. Pb isotope ratios for wall painting samples from Ephesus, ore samples from the Karaburun 50 Peninsula and reference data for A Italy, Turkey, Greece, and Spain, B as well as with Balkan reference 51 data from the AAcP database. 52 53 (2-column fitting image) 54 Page 10 of 10 55 56 57 58 Figure 7. Hg isotope ratio data (A Δ199Hg vs δ202Hg, B Δ199Hg vs Δ201Hg) for wall painting samples from 59 Ephesus, ore samples from the Karaburun Peninsula, and cinnabar ore samples from Italy (Pribil et al., 60 2020), Spain (Gray et al., 2013), China (Yin et al., 2013; Ni et al., 2022), and USA (Stetson et al., 2009). 61 (2-column fitting image) 62