Examination and Analysis of Etruscan Wall Paintings at Caere, Italy

bs_bs_banner Archaeometry ••, •• (2017) ••–•• doi: 10.1111/arcm.12304 E X A M I N AT I O N A N D A N A LY S I S O F E T R U S C A N WA L L PA I N T I N G S AT C A E R E , I TA LY * B. KLEMPAN† Department of Art History and Art Conservation, Queen’s University, Kingston, ON K7L 3N6, Canada K. HELWIG Canadian Conservation Institute, 1030 Innes Road, Ottawa, ON K1B 4S7, Canada and F. COLIVICCHI Department of Classics, Queen’s University, Kingston, ON K7L 3N6, Canada In 1983, the excavation of the ancient city of Caere near Rome, which became a UNESCO World Heritage Site in 2004, led to the unearthing of an underground sanctuary dating from the early third century BC. Known as the ‘Hypogaeum of Clepsina’, it consists of an underground room decorated with frescoes, drawings and inscriptions. The initiation of a new excavation campaign at Caere in 2012 provided the opportunity to study this rare example of Etruscan wall paintings of a non-sepulchral nature from the Hellenistic era. The paintings were documented, photographed and samples were removed for analysis using a combination of scientific techniques. The observation that at least some of the painting was done on wet plaster supports the conclusion that the frescos are contemporaneous to the plastering of the walls in the hypogaeum in approximately 273 BC. The plaster is lime-based and aggregate materials include potassium feldspar, clinopyroxene, quartz, hematite and magnetite. No organic binding medium was identified in the paint, suggesting a lime water binder. Paint materials identified are Egyptian blue, red iron oxide (hematite), yellow iron oxide (goethite), charcoal black, quartz, silicates and calcite. All the paint samples show a mineral accretion layer at the upper surface, caused by recrystallization of salts. KEYWORDS: ETRUSCAN WALL PAINTING, FRESCO, LIME, PIGMENTS, SEM, FTIR, Py–GC–MS, XRD, PLM, CROSS-SECTIONS INTRODUCTION The so-called ‘Hypogaeum of Clepsina’ is located in the centre of the plateau of the ancient city of Caere, one of the most important sites of ancient Etruria, located ~45 km north-west of Rome. The focus of the complex, a religious compound for chthonic rituals that may have been the symbolic centre of the urban space, is an underground room with two large windows that open on to a deep square court with a complex system of access and circulation (Cristofani and Gregori 1987; Torelli 2000; Colivicchi 2014; Colivicchi et al. 2016). This monument is directly connected to the political transformation of Caere into a Roman community. On a wall of the room is the inscription of a prominent figure of the Roman political scene, C. Genucius Clepsina, consul in 276 and 270 BC. He was the first Prefect of the newly established Praefectura of Caere, sent by Rome to dispense justice and administer the new addition to the Roman territory. The final incorporation of Caere into the Roman state with the establishment of the Praefectura Caeritum is not well explained by ancient authors, but the event can be dated with reasonable approximation to 273 BC (Sordi 1960, 123–34; Humbert 1972; Torelli 2000, 154–5). The text *Received 5 August 2016; accepted 21 December 2016 †Corresponding author: email klempanb@queensu.ca © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford Reproduced with the permission of the Minister of Canadian Heritage. B. Klempan, K. Helwig and F. Colivicchi 2 of the inscription of Clepsina was carved on wet plaster during the construction of the building and dates it to about 273 BC with a precision that is absolutely exceptional for such an early period. The focus of the underground room was a rectangular niche, which was decorated with technically accomplished wall paintings, a palm tree on the two side walls and almost indecipherable figures on the back wall. The paintings of the hypogaeum have been discussed from the point of view of their symbolic meaning (Cristofani and Gregori 1987; Torelli 2000), but a technical study was sorely needed, especially given their exceptional importance for the study of Etruscan and Roman wall painting of this period. Etruscan painting of the third century BC is almost entirely represented by painted tombs, and its chronology has been established on the ground of stylistic sequences and the archaeological dating of associated artefacts, such as the sarcophagi of Tarquinia (Colonna 1984). The Hypogaeum of Clepsina is especially important because it is a public building independently dated with extreme precision by its connection to historical events and figures. The site of the hypogaeum has already provided valuable new data on Etruscan paintings, this time on its early phase. Previous research near the entrance of the monument uncovered remains of a building of the early seventh century BC with walls decorated by painted geometric patterns. Fragments of painted plaster were analysed by scientists at the University of Perugia. Using a variety of analytical techniques, they determined that the painting substrate consisted of a mixture of powdered volcanic stone and calcareous rock. Orange, red and white were the only colours employed to create the simple designs on the surface. All of the pigments were identified as mineral in nature and appeared to have been applied in a mixture of lime water (Miliani et al. 2003). Analysis of the wall paintings in the niche, described in this paper, adds further information about this important archaeological site. EXAMINATION AND DOCUMENTATION OF THE WALL PAINTINGS A condition survey, including standard and high-resolution digital photographic documentation and sampling of the three wall paintings at Caere was undertaken in June 2012. The three wall paintings are located in a small rectangular niche within the hypogaeum (Fig. 1). The dimensions of the paintings are as follows: north-west wall, H = 171.0 cm at highest point, W = 90.0 cm at widest point; north-east wall, H = 49.0 cm at highest point, W = 68.0 cm at widest point; southeast wall, H = 172.0 cm at highest point, W = 89.5 cm at widest point. Painting technique The two larger paintings of palm trees on the north-west and south-east walls were painted in a direct manner on what may have been a wet, or at least still soft, plaster surface, similar to the buon fresco technique, where pigment is applied to a wet plaster surface. This is supported by the observation of unintentional indentations in the plaster layer (Fig. 2). Detailed examination of these indentations clearly showed that paint did not fill the indentation as one might expect by placing wet paint over a hardened gouge in the plaster. Instead, the newly applied paint was pushed back with the soft plaster layer when the damage occurred. This observation, along with the continuity of the plaster applied to all surfaces in the hypogaeum, strongly suggests that the paintings were done at the same time as the plastering of the walls. It is difficult to say with certainty, based on limited examples of indentations in painted areas, whether the entire rendition of palm trees was done on wet plaster. The paintings could be a combination of working on wet plaster and dried plaster. © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 3 Figure 1 The niche with three wall paintings, showing the overall photograph (north-east wall) in the centre (b), and details on the left (a, north-west wall) and right (c, south-east wall) (photographs by B. Klempan). Figure 2 A detail of the south-east wall painting with a textured plaster surface and indentations in the wet plaster (photograph by B. Klempan). On several areas of the painted walls, there was a visible, distinctive brush pattern on the surface of the plaster. This pattern may be the result of applying a final layer of plaster with a stiff bristle brush to even out the surface (Fig. 2). © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• 4 B. Klempan, K. Helwig and F. Colivicchi There is clear evidence of a layered build-up of paint on the palm trees done in the fresco secco technique or ‘dry technique’, where pigments are mixed with a binder such as lime water and applied to a dry plaster surface. The palm fronds, for example, on the south-east wall were first painted with a thin application of red paint, which is directly on top of the plaster layer, followed by a thin light green paint application on top of the red paint. Details on these fronds were undertaken first with a darker green paint and lastly a deep blue paint (Fig. 3 (a)). The trunks of the palms have detailed, artistic renderings of a textured surface and are composed mainly of red, yellow and dark black–brown paint. The dark black–brown paint was applied over the other layers to produce a cross-hatched texture to the trunk (Fig. 3 (b)). This type of layering structure clearly indicates a fresco secco technique. The small painting on the back wall (north-east wall), although severely water damaged, appeared to show a less sophisticated style of painting with a restricted colour palette. Condition The hypogaeum is remarkably well preserved considering the ancient nature of the site. The paintings, apart from the small painting above the back wall (north-east wall), which had sustained severe water damage (Fig. 1 (b)), were in generally good condition apart from some large losses into the tufa base and some abraded areas of paint (Figs 1 (a) and 1 (c)). The south-east wall had some damages that appeared as deep gouges in the painting’s surface. Smaller gouges were seen throughout all three paintings and they may have been caused by the removal of iron nails from the wall for reuse. Other forms of damage included narrow cracks in the plaster and paint layers, and beaded deposits of efflorescence were observed exuding from some cracks. Major damages were measured, described and photographed for future comparative analysis. Environmental conditions within the hypogaeum were monitored during site work using two HOBO® temperature and humidity data loggers. During the period from 4 to 28 June 2012, readings were taken every 5 min for a total of 6947 readings. The average temperature was Figure 3 Images of paint application: (a) a detail of the palm fronds on south-east wall palm tree; (b) a detail of the palm trunk on the north-west wall (photographs by A. Gabov). © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 5 14.39°C; the average relative humidity was 99.40% and the average dew point was 14.31°C. Droplets of water on the ceiling and wet walls confirmed the high humidity conditions within the hypogaeum. This short span of environmental monitoring was done when the door to the hypogaeum was open during the day for work and visitors and closed at night. It is probable that the humidity was not as high in the hypogaeum during off-season periods when the door was closed and access restricted. All of the walls in the hypogaeum, including those with paintings, had a slight sheen to their surface, due to the presence of a thin, mineral accretion layer. There were also areas of a more opaque white layer producing a blanched appearance on the surface. The surface accretion, blanching and efflorescence are the result of the uncontrolled environmental conditions. Sampling Thirty samples were removed from the hypogaeum, with the majority of samples taken from the wall paintings. Each sample was taken with a new, stainless steel scalpel blade and precise measurements were taken to record sample locations. The removal of samples proved to be very difficult as the paint and walls were extremely hard and not at all friable. This may explain why the paintings are in such good condition after so many centuries underground. Of the 30 samples that were collected, a subset of 13, which included samples of the plaster and all the major colours on the paintings, were chosen for analysis. The goals of the analysis were to determine the composition and stratigraphy of the plaster and paint layers, to establish whether an organic binding medium had been used, and to determine the composition of the surface accretion layer. A description of the samples that were analysed and their locations is shown in the first column of Table 1. METHODS OF ANALYSIS Fragments of the paint and plaster samples were analysed using a combination of scanning electron microscopy–energy dispersive spectroscopy (SEM–EDS), Raman spectroscopy, Fourier transform infrared spectroscopy (FT–IR), pyrolysis–gas chromatography – mass spectrometry (Py–GC–MS) and polarized light microscopy (PLM). For each technique, analysis was undertaken on a representative sub-fragment of the overall sample. SEM–EDS was performed using an Hitachi S-3500 N VP SEM integrated with an Oxford Inca X-act analytical silicon drift X-ray detector and an Inca Energy + X-ray microanalysis system. The SEM was operated at an accelerating voltage of 20 kV in high-vacuum mode, using a backscattered electron detector. Using this technique, elemental analysis of volumes down to a few cubic micrometres can be obtained for elements from boron (B) to uranium (U) in the periodic table at a level of approximately 0.1–1% or greater. The cross-sections were carbon coated prior to analysis to ensure conductivity. Fragments of paint and plaster were adhered to a carbon planchet with carbon tape and analysed directly by SEM–EDS without carbon coating. Raman spectra were collected with a Bruker Senterra dispersive Raman microscope, using a 100 mW diode laser with a wavelength of 785 nm. A 50× objective lens was used to produce an analysis area of approximately 2 μm in diameter. The laser power at the sample was less than 1 mW. For FT–IR, small fragments of the samples were mounted directly, without pre-treatment, on a low-pressure diamond anvil microsample cell and analysed using a Bruker Hyperion 2000 microscope interfaced to a Tensor 27 spectrometer. The samples were analysed in transmission mode by co-adding 200 scans and using a 4 cm 1 resolution. Spectra were collected from © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• B. Klempan, K. Helwig and F. Colivicchi 6 Table 1 Results of analysis Sample description and location (in cm, as measured from top (T), right (R) and left (L) edges of wall) Pigments identified (FT–IR, Raman, XRD and PLM) SEM–EDS* 1. Upper white plaster; north-west wall: T = 54.0; R = 77.0 White matrix: calcite, silicates Aggregate minerals: hematite, quartz, potassium feldspar, clinopyroxene Ca, O, Si, C, Mg, Al, (Fe) 2. Lower beige plaster and small amount of upper plaster; north-west wall: T = 49.0; R = 38.0 Beige matrix: calcite, aragonite, silicates Aggregate minerals: potassium feldspar, magnetite, clinopyroxene, quartz, probably iron and manganese oxides Ca, Si, O, Al, C, Mg, K, (Fe) 3. Blue, fourth complete frond in palm; north-west wall: T = 36.5; R = 54.0 Egyptian blue, charcoal black, calcite, small amount yellow iron oxide Si, O, Ca, Cu, C 4. Blue, right side near ochre colour; north-east wall: T = 16; R = 4.0 Egyptian blue, charcoal black, calcite Ca, O, Si, C, (Cu, Fe, Al) 5. Green and plaster below, palm, fifth large frond; north-west wall: T = 35.5; R = 153.4 Egyptian blue, calcite, charcoal black, yellow iron oxide, small amount red iron oxide Ca, Si, C, O, Cu, (Al, Fe) blue particle: Si, Ca, O, Cu, C, (Cr) 6. Yellow with orange below, palm trunk north-west wall: T = 63.0; R = 46.0 Yellow iron oxide (goethite), red iron oxide (hematite), calcite, silicates Ca, O, Fe, C, Si, (Mg, Al) 7. Pale red and plaster below, palm trunk; north-west wall: T = 80.5; R = 49.0 Calcite, red iron oxide (hematite), silicates Ca, O, Si, C, (Mg, Al, P, Fe) 8. Orange–red, in upper right; north-west wall: T = 10.0; R = 5.5 Yellow and red iron oxide (goethite, hematite), calcite, silicates O, Si, Fe, Ca, C, (Al) 9. Brownish purple paint and plaster below, palm trunk; south-east wall: T = 7.5; L = 49.5 Calcite, red iron oxide (hematite), silicates, charcoal black, yellow iron oxide Ca, C, O, (Si, Al, Fe) 10. Brown and plaster below, end of damage on frond; south-east wall: T = 41.0; L = 53.5 Red iron oxide (hematite), yellow iron oxide (goethite), Egyptian blue, charcoal black, calcite Ca, C, O, Si, Fe, Al, (Cu, Mg, K) 11. Black, inside palm trunk; south-east wall: T = 47.5; L = 44.5 Charcoal black, Egyptian blue, yellow iron oxide, calcite, small amount red iron oxide Ca, O, C, Si, (Fe, Al) 12. Black and plaster below, faint figures; north-east wall: T = 18.0; L = 25.0 Charcoal black Ca, C, O, (Mg, Al, Si, S) 13. Efflorescence; north-west wall: T = 105.5; R = 79.0 Calcite, silica Ca, C, O, Si, (Mg, Al, P, S) *Based on relative peak height, the chemical elements are qualitatively categorized as major (underlined), minor (normal typeface) and trace (in parentheses). The presence of carbon in all samples is due in part to the carbon tape used to mount the samples. 4000 to 430 cm 1 using a wide-band MCT detector. Spectra of the samples were compared to published reference spectra. PLM was undertaken on fragments of the pigmented samples using a Leica DMRX polarized light microscope. The samples were mounted in Cargille Meltmount (n = 1.66) and particles characterized on the basis of their optical and morphological properties (size, © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 7 shape, colour and birefringence). The dispersed pigment particles were observed using a total magnification of 400. Analysis was performed by Py–GC–MS after derivatization using tetramethyl ammonium hydroxide (TMAH, 2.5% in methanol). For each analysis, a small fragment of the sample was placed in a micro-vial (Agilent Technologies, part no. 5190–3187) with 1.6 μL of tetramethylammonium hydroxide (TMAH, Supelco, Bellafonte, PA) in methanol (1:25). The vial was inserted into a thermal separation probe (TSP, Agilent Technologies, Inc., Palo Alto, CA) installed in a multimode inlet on an Agilent 7890A gas chromatograph interfaced to a 5975C mass-selective detector. The multimode injector with TSP was operated in splitless or split mode depending on the sample size and ramped from 50°C to 450°C, at a rate of 900°C min 1 to perform the pyrolysis. The final temperature was held constant for three minutes and then decreased to 250°C at a rate of 50°C min 1. For the gas chromatographic separation, a Phenomenex ZB-Semivolatiles fused silica column (30 m × 0.25 mm i.d., 0.25 μm film thickness; Phenomenex Inc., Torrance, CA) was used. Ultra-high-purity helium carrier gas was used with a constant flow of 1.2 mL min 1. The oven was programmed from 40°C to 200°C at 10°C min 1 and from 200°C to 300°C at 6°C min 1, with a hold time of 20 min (52.67 min run time). The MS was operated in EI positive mode (70 eV). The MS transfer line temperature was 280°C; the MS ion source was held at 230°C and the MS quadrupole at 150°C. The MS was run in scan mode from 50 to 550 amu (5–25 min), 50–750 amu (25–30 min) and 50–800 amu (35–63 min). Data were processed using Agilent ChemStation software (v.E.02.02). Eight of the 13 samples were prepared as cross-sections. A small fragment of each sample was mounted in polyester resin and ground and polished with silicon carbide abrasive papers to expose the stratigraphy. The cross-sections were observed by incident light microscopy, using a Leica DMRX polarizing light microscope. Layers of interest in the cross-sections were analysed using scanning electron microscopy–energy dispersive spectrometry (SEM–EDS) and Raman spectroscopy. Both incident light and backscattered electron images of the cross-sections were collected. RESULTS AND DISCUSSION The results of the analysis are shown in Table 1 and discussed below. Plaster The wall paintings were executed on a substrate composed of a layer of coarse, beige plaster followed by an off-white finishing plaster with a finer texture. The overall thickness of the coarse and fine plaster was not determined, since none of the samples contained the complete stratigraphy of the rock substrate and plaster layers. A cross-section of the plaster layers, showing both the coarse beige plaster and the finishing plaster, is illustrated in Figure 4. The coarse plaster has a beige matrix with black, transparent and brown aggregates. The matrix is composed of calcium carbonate and fine-grained silicates, a typical composition for lime plaster. Several of the transparent aggregate particles were found to be potassium feldspar. One dark aggregate was identified as clinopyroxene, while another was found to be magnetite. A deep brown–black particle contained quartz and, based on the chemical elements, probably manganese and iron oxides. The aggregate materials identified are similar to those identified by Miliani et al. (2003) and are consistent with the use of local crushed stone. A thinner layer of finishing plaster was applied on top of the coarse plaster. Although the finishing plaster has a similar calcium carbonate matrix to the coarse plaster, it does not include © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• 8 Figure 4 B. Klempan, K. Helwig and F. Colivicchi A cross-section showing coarse, beige plaster, followed by part of the upper, finishing plaster: incident light. the large aggregate minerals and visually appears much finer and lighter in colour. Some relatively finely divided transparent, black, brown and red particles are present and give the plaster a pale ochre appearance on the wall. Some of these aggregate particles are visible in Figure 4, and were also seen in the fine plaster layer below the paint in many of the cross-sections from painted areas. Analysis of one of the transparent inclusions in the fine plaster showed that it is potassium feldspar. A translucent, brown inclusion was identified as clinopyroxene and a red inclusion was found to contain quartz and hematite. Paint All the paint samples analysed contain calcium carbonate mixed with coloured pigments. The components identified in the paint samples are listed in Table 1. Using FTIR, no organic binding medium was identified in any of the paints. Four of the samples (6, 7, 9 and 11) were subsequently analysed using Py–GC–MS to determine if trace residues of organic compounds could be detected. No organic compounds were identified, suggesting that either the organic binding medium had completely deteriorated or that no organic medium was used. Since Py–GC–MS is a very sensitive technique, even for highly degraded proteins and lipids, it seems more likely that no organic binder was used on the wall paintings. As described earlier, the visual examination of the paintings suggested that parts of the palm trees were painted on wet plaster, but that some of the details were added in fresco secco. That no organic binder was found in the paint from the hypogaeum suggests that the fresco secco application was done with lime water. The lack of an organic binder is of interest since in several previous studies of Etruscan wall paintings, egg was identified as a binding material (Colombini et al. 2003; Pallecchi et al. 2009). The following pigments and accessory materials were identified in samples from the wall paintings in the hypogaeum: red iron oxide (hematite), yellow iron oxide (goethite), Egyptian blue, charcoal black, quartz, silicates and calcium carbonate (calcite form). The pigments used on the wall paintings are compatible with lime (Mora et al. 1984) and have all been previously identified in Etruscan wall paintings and polychromes (Colombini et al. 2003; Miliani et al. 2003; Sodo et al. 2008; Pallecchi et al. 2009; Giachi et al. 2015; Brøns et al. 2016). What is particularly interesting about the wall paintings in the hypogaeum is the range of pigment mixtures that were used to produce different shades. Colours such as green, brown and black were © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 9 found to contain mixtures of all the pigments on the artist’s palette; by varying the proportions of the colourants, a wide range of different shades was obtained. Blue Blue paint samples from the north-west and north-east walls both contain Egyptian blue, charcoal black and calcium carbonate. Egyptian blue, coloured by a calcium copper silicate (cuprorivaite structure), is is considered to be the first synthetic pigment. It was used in ancient Egypt from the third millennium BC and disappeared from use after the fall of the Roman Empire (Reiderer 1997; Daniels et al. 2004). Egyptian blue pigment, which could have been imported or synthesized locally, is known to have been used by Etruscan artists (Colombini et al. 2003; Sodo et al. 2008; Pallecchi et al. 2009; Giachi et al. 2015; Brøns et al. 2016). It is made by heating a calcium compound (such as limestone), silica (usually in the form of sand), a source of copper (such as malachite, copper or bronze fragments) and a small amount of alkali to temperatures around 850–1000°C (Daniels et al. 2004). Traces of chromium were identified in some of the Egyptian blue particles analysed in the paint from the hypogaeum. The presence of specific trace elements is related to the nature of the starting materials used for the synthesis (Reiderer 1997). Yellow Yellow paint from a highlight on the palm trunk on the north-west wall was found to be pigmented with yellow iron oxide (goethite), along with smaller amounts of red iron oxide (hematite) and calcium carbonate. Iron oxide pigments were among the first pigments known and written about. They were highly valued and widely used during the classical period (Helwig 2007). Red and yellow iron oxides have been previously identified on Etruscan wall paintings, including those at Caere from the early seventh century BC building near the entrance of the hypogaeum (Miliani et al. 2003). A cross-section of the paint from the palm trunk, illustrated in Figure 5, shows the superposition of two layers of paint—the pale red–brown of the palm trunk followed by the bright yellow highlight on the diamond pattern of the bark. Consistent with the visual examination, the presence of multiple paint layers with a clear interface indicates that the Figure 5 A cross-section from the palm trunk on the north-west wall. The cross-section shows the superposition of the bright yellow highlight on the diamond pattern of the bark on the pale red–brown paint of the palm trunk. The upper layer in the cross-section is the white mineral accretion, present in all areas of the wall paintings that were examined. © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• 10 B. Klempan, K. Helwig and F. Colivicchi yellow paint was done in fresco secco, after the paint beneath had dried. The cross-section also shows a white mineral accretion on top of the yellow paint. Red Several different shades of red were investigated. Colours on the wall paintings described as pale red, orange–red and brownish purple were all found to be based on red iron oxide with the hematite structure. The paint described as brownish purple contains primarily red iron oxide, but also some charcoal black, as well as silicates, calcium carbonate and a small amount of yellow iron oxide. The darker, browner tone of this paint is mainly due to the addition of charcoal black to the red oxide. This differs from the dark red painted plaster from the early seventh century BC building near the entrance of the hypogaeum, which was found to be a mixture of two forms of iron oxide (hematite and maghemite) along with a black manganese oxide (pyrolusite) (Miliani et al. 2003). In classical times, different shades of red were sometimes produced by using hematite of different particle size—since the typical red colour of hematite becomes purple as the particle size increases (Helwig 2007). This does not appear to be the case in these wall paintings, where, instead, the darker shade was produced by mixing charcoal black into the red paint. The pale red paint contains only red iron oxide and has been mixed with a higher proportion of calcium carbonate to lighten the tone. The orange–red paint contains both red and yellow iron oxides (hematite and goethite) as well as silicates and calcium carbonate. Green Although green pigments were available during the classical period, the green colour on wall paintings in the hypogaeum was created with a mixture of blue and yellow. Analysis of green paint from the north-west wall showed that it was produced with Egyptian blue mixed with yellow iron oxide (goethite). Calcium carbonate and small amounts of red iron oxide (hematite) and charcoal black were also added. Brown Brown paint from a palm frond from the south-east wall is a mixture of red and yellow iron oxides (hematite and goethite), Egyptian blue, charcoal black and calcite. Similar brown colours made from mixtures of iron oxides with Egyptian blue have been previously identified on Greek and Roman wall paintings (Helwig 2007 and references therein). Black Black paint from two different areas was examined; black paint from the palm trunk on the southeast wall and black paint from the barely discernible figures on the north-east wall above the tunnel. Charocal black was identified in both areas. The black pigment particles showed the typical coarse, angular morphology of plant-based chars and contained fragments with residual cellular structure (Winter and Fitzhugh 2007). As described previously, the paintings on the north-east wall are stylistically different and may not have been painted at the same time as the palm trees on the north-west and south-east walls. This is supported by the microscopic differences seen in the paint. The black paint from the palm trunk is a complex mixture of pigments; it contains charcoal black, mixed with red and yellow © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 11 iron oxides and Egyptian blue. The black from the faint figures on the north-east wall, on the other hand, is a thin layer of charcoal black, unmixed with other pigments. Mineral accretions and efflorescence All the cross-section samples were found to have a compact, white mineral accretion on the top surface above the paint (Fig. 6). This corresponds to the observation that the paintings showed an overall sheen, with thicker blanching in specific areas. The composition of the accretion is the same over the surface of the wall paintings. It contains primarily finely precipitated calcium carbonate with silica in some areas. Figure 6 A cross-section showing brown paint with the surface accretion above: (a) incident light; (b) a backscattered electron image. The brown paint is a complex mixture containing red iron oxide (hematite), yellow iron oxide (goethite), Egyptian blue, charcoal black and calcite. The backscattered electron image shows the compact, uniform nature of the accretion, which is composed primarily of finely precipitated calcium carbonate. © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• 12 B. Klempan, K. Helwig and F. Colivicchi Similar mineral layers have been noted on other Etruscan wall paintings (Pallecchi et al. 2009). This hard, glassy precipitate has been produced by the slow aqueous leaching of calcium and silicon from the environment, followed by precipitation at the surface as the water evaporated (Mora et al. 1984). There are a number of sources of moisture in the chamber that are probably involved: for example, condensation of humid air on the surface of the walls, ground water penetration through the walls and rising damp. This type of accretion is hard and insoluble so is difficult to remove, but is usually less harmful than sulphate salts, which can cause significant expansion and disintegration of the wall surface (Mora et al. 1984). Because the accretion is thin, it does not obscure the paintings to a great extent, and in fact functions as a protective coating. By leaving the surface accretion in place and controlling the environment, the surface of the paintings would be expected to remain stable. As described earlier, the wall paintings also show local regions of efflorescence, exuding from cracks. This efflorescence has a similar composition and source as the superficial mineral accretion. CONCLUSIONS The examination and analysis of wall paintings located in the Hypogaeum of Clepsina has revealed an ancient and fascinating past. The paintings of palm trees on two walls of the hypogaeum show a skilled and direct handling of paint and reveal a sophisticated approach to wall decoration for such early Etruscan works. There is a complex build-up of paint layers on the palm trees to add the illusion of texture and structure to the depiction. The smaller painting on the north-east wall is difficult to interpret due to its poor overall condition. The technical examination of these paintings has helped define a history of their survival and record their current condition that can be used for comparative analysis in the future. The detailed recording of surface characteristics through high-resolution digital photography and field note recording captured areas of interest such as: exact areas where chunks of tufa and plaster had been dug out of the walls; locations where iron nails had been removed from the walls; blanching of the surface; cracks in the plaster; and indentations in both paint and plaster. The observation that at least some of the palm trees were painted on wet plaster strongly supports the conclusion that the incised inscriptions and paintings are contemporaneous to the plastering of the walls in the hypogaeum in approximately 273 BC. The scientific analysis showed that the wall paintings were executed on a lime plaster substrate composed of a coarse, beige application followed by white finishing plaster. The composition of the aggregate particles is consistent with the use of local crushed stone. The following materials were used to create the painted design: red and yellow iron oxides, Egyptian blue, charcoal black, quartz, silicates and calcite. Although the palette is limited, a wide range of pigment mixtures was used to produce delicate and subtle colour variations. No organic binding medium was identified in the paint, suggesting that the only binder was lime water. Based on the visual examination and the analysis of cross-sections, it can be concluded that parts of the palm trees were painted when the plaster was still wet, but that subsequent paint was added in a fresco secco technique. A white, mineral accretion, composed primarily of finely precipitated calcium carbonate, is present on the surface of the wall paintings. This thin surface accretion does not obscure the paintings to a great extent, and in fact functions as a protective coating. By leaving the mineral accretion in place and controlling the environment, the surface of the paintings would be expected to remain stable. The recording and observation of the environmental conditions within the hypogaeum is fundamental in comprehending the physical survival of these paintings in an © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–•• Examination and analysis of Etruscan wall paintings at Caere, Italy 13 underground chamber. Data loggers currently in the chamber should provide readings for future comparison. ACKNOWLEDGEMENTS We are grateful to Jennifer Poulin, Dominique Duguay and Clémentine Mansas for assistance with the scientific analysis. We also thank Krysia Spirydowicz for her support during the project, as well as Alexander Gabov and Anna Weiss for their assistance. REFERENCES Brøns, C., Hedegaard, S. S., and Sargent, M. L., 2016, Painted faces: investigations of polychromy on Etruscan antefixes in the Ny Carlsberg Glyptotek, Etruscan Studies, 19, 23–67. Colivicchi, F., 2014, The Mundus of Caere and early Etruscan urbanization, in Urban dreams and realities in antiquity: remains and representations of the ancient city (ed. A. Kemezis), 46–68, Brill, Leiden. Colivicchi, F., Gregori, G. L., Lanza, M., Lepone, A., Scalici, M., Trentacoste, A., and Zaccagnino, C., 2016, New excavations in the urban area of Caere, Mouseion: Journal of the Classical Association of Canada, 13, 359–450. Colombini, M. A., Giachi, G., Pallecchi, P., and Ribechini, P., 2003, Tecniche pittoriche utilizzate nelle tombe etrusche di Chiusi e Sarteano, in Pittura etrusca: problemi e prospettive: atti del convegno, Sarteano, Teatro comunale degli Arrischianti, 26 ottobre 2001, Chiusi, Teatro comunale Mascagni, 27 ottobre 2001 (ed. A. Minetti), 162–7, Amministrazione provinciale di Siena, Siena. Colonna, A., 1984, Per una cronologia della pittura etrusca di età ellenistica, Dialoghi di Archeologia, 2, 1–24. Cristofani, M., and Gregori, G. L., 1987, Di un complesso sotterraneo scoperto nell’area urbana di Caere, Prospettiva, 49, 2–14. Daniels, V., Stacey, R., and Middleton, A., 2004, The blackening of paint containing Egyptian blue, Studies in Conservation, 49, 217–30. Giachi, G., Pallecchi, P., Marras, L., and Palleschi, V., 2015, Dal colore dipinto al colore ritrovato: le pitture degli ipogei di Chiusi, in La Tomba de Colle nella passeggiata archeologica a Chiusi (eds. M. Salvini, G. Paolucci, and P. Pallecchi), 39–48, Quasar, Roma. Helwig, K., 2007, Iron oxide pigments: natural and synthetic, in Artists’ pigments: a handbook of their history and characteristics (ed. E. W. Fitzhugh), 39–109, National Gallery of Art, Washington, DC. Humbert, M., 1972, L’incorporation de Caere dans la civitas Romana, Mélanges de l’Ecole française de Rome Antiquité, 84(1), 231–68. Miliani, C., Rosi, F., Borgia, I., Perugini, D., Cartechini, L., Poli, G., Brunetti, B. G., and Sgamellottio, A., 2003, Studio della tecnica pittorica dei dipinti murali arcaici dell’area dell’ipogeo di C. Genucius Clepsina, Science and Technology for Cultural Heritage, 12, 91–7. Mora, P., Mora, L., and Philippot, P., 1984, Conservation of wall paintings, Butterworths, London, 56–104, 165–215. Pallecchi, P., Giachi, G., Colombini, M. P., Modugno, F., and Ribechini, E., 2009, The painting of the Etruscan ‘Tomba della Quadriga Infernale’ (4th century BC), in Sarteano (Siena, Italy): technical features, Journal of Archaeological Science, 36, 2635–42. Reiderer, J., 1997, Egyptian blue, in Artists’ pigments: a handbook of their history and characteristics (ed. R. L. Feller), 23–45, National Gallery of Art, Washington, DC. Sodo, A., Artioli, D., Botti, A., De Palma, G., Giovagnoli, A., Mariottini, M., Paradisi, A., Polidoro, C., and Ricci, M. A., 2008, The colours of Etruscan painting: a study of the Tomba dell’Orco in the necropolis of Tarquinia, Journal of Raman Spectroscopy, 39, 1035–41. Sordi, M., 1960, I rapporti romano-ceriti e l’origine della civitas sine suffragio, ‘L’Erma’ di Bretschneider, Roma. Torelli, M., 2000, Genucio(s) Clousino(s) prai(fectos): la fondazione della praefectura Caeritum, in The Roman middle Republic: politics, religion, and historiography, c.400–133 B.C.: papers from a conference at the Institutum Romanum Finlandiae, September 11–12, 1998 (ed. C. Bruun), 141–76, Acta Instituti Romani Finlandiae, Rome. Winter, J., and Fitzhugh, E. W., 2007, Pigments based on carbon, in Artists’ pigments: a handbook of their history and characteristics (ed. E. W. Fitzhugh), 1–37, National Gallery of Art, Washington, DC. © 2017 Her Majesty the Queen in Right of Canada. Archaeometry © 2017 University of Oxford, Archaeometry ••, •• (2017) ••–••