Revealing the unseen: advanced digital imaging and spectroscopy for the documentation and non-invasive analysis of a 16th-century Cypriot composite icon

Abstract

The present study focuses on an icon known as the Panagia Valieriotissa at the Church of Saint Nicholas in Palodeia, Cyprus. The icon is unusual as it contains two depictions of the Virgin Mary and Child. The main aim of this study is to analyze the material characteristics of the icon and the relation of the two panels in order to determine if the icon belongs to a distinct category of paintings known as “composite icons” – icons consisting of two separate panels, one inserted into another. Icons of this unusual type, first described by Panayotis Vocotopoulos, appear to be quite numerous on Cyprus and date mostly to the sixteenth century. Results show that the icon can be included within this group, consequently enriching our knowledge on the materials, technologies, and techniques used in the creation of composite icons. Driven by art historical inquiries, the study’s methodological approach was based on the integration of digital imaging techniques with non-invasive and non-destructive material analyses pursued by the Andreas Pittas Art Characterization Laboratories (APAC Labs) of the Cyprus Institute. Spectroscopic techniques, digital microscopy and advanced digital imaging methods produced a broad range of data which were analyzed and interpreted from both an art-historical and a conservation perspective towards a better understanding of the materiality of the Palodeia icon and its subsequent history and use. The imaging and spectroscopic methods used provided new information on the materials and techniques used by the painters as well as on later undocumented episodes of restoration or repainting. Overall, the results showed that the icon indeed consists of two separate icons – a larger panel and a smaller, re-used icon inserted into the former. Moreover, some important distinctive features of the inset icon, invisible to the naked eye, reveal its possible previous use and dating. The study brought to light the great potential of non-invasive methods in the revealing and analysis of unknown aspects of composite icons, thus shedding light on this fascinating phenomenon, but it can also interest a broader group of researchers into the exciting new possibilities as well as the limitations of non-invasive methods for the study of works of art and heritage objects.

Introduction

This paper presents the results of the study of an icon, known as the Panagia Valieriotissa, that is still venerated at the Church of Saint Nicholas in Palodeia, Cyprus. The icon has an unusual form and iconography as it contains two images of the Virgin Mary and Child, suggesting that it could belong to the distinct group of “composite icons.” This particular group concerns icons consisting of two separate panels, often created at different times, one inserted into another. The term was coined by Panagiotis Vocotopoulos: the first scholar to note and describe this particular type of Byzantine and post-Byzantine icons. His article published in Greek and English attempted to identify such icons from various regions, spanning a wide timeframe from the thirteenth to the nineteenth centuries. The study remains a valuable contribution in the study and cataloguing of composite icons, being considered the most important publication describing the phenomenon of the re-use of icons by inserting them into larger panels [1, 2]. Cyprus preserves numerous examples of this icon type [3, 4]. It is important to note that the icon from Palodeia was not included in Vocotopolos’ study. In fact, the icon’s current state of preservation makes it difficult to establish with certainty whether the object indeed consists of two separate panels, one inserted into another, or is rather a single panel with a painted niche – and thus an imitation of the form. In such cases, a technical examination, especially with the application of X-ray radiography, is particularly useful, usually providing conclusive results. For example, the technical study of a painting by Botticelli, which accompanied the recent auction of the artwork at Sotheby’s, included an older painting inserted into the master panel and provided valuable information about the practice of the re-use of paintings through their insertion into larger panels[5]. Unfortunately, hitherto no comparable analyses have been conducted on any composite icon, even though a broad range of technological and analytical examinations of icons have been published in recent years [6,7,8,9,10,11,12,13,14,15]. Among these approaches, non-invasive methods appear to be very effective in the study of the materiality and technique of icon painting [16,17,18,19,20]. Within this methodological framework, the application of non-invasive analytical techniques can help answer the question as to whether the Palodeia icon is an example of a composite icon. The present research is the first attempt to use these methods to study the materiality of composite icons, focusing on an important but little-known example, namely the icon of the Panagia Valieriotissa. Moreover, this study seeks to extend our knowledge on the materials and techniques used for the creation of the icon as well as its subsequent history, thus providing a better understanding of its unusual form and iconography.

Panagia Valieriotissa icon, its complex form and history

The icon analyzed here is a 16th-century icon (109.5 × 67 × 4 cm) from the village of Palodeia (Fig. 1), located approximately ten kilometers northwest of the port city of Limassol. It depicts the Virgin Mary and Child and is accompanied by the inscription ΜΗ(ΤΗ)Ρ / Θ(Ε)ΟΥ / Η ΒΑΛΗ/ΕΡΗΟΤΗ/ΣΣΑ, providing evidence of the dedication to the Panagia Valieriotissa. A smaller icon (23.5 × 16.5 cm), located in the center of the panel (Fig. 2), also depicts the Virgin Mary with Christ on her left arm. The icon’s current form and state of preservation reflects its history and previous uses. The icon comes from the nearby Church of Panagia Valieriotissa (or Vakeriotissa as the locals called it), but it was moved at an unknown date to the Church of Saint Nicholas in Palodeia and is presently located in the church’s iconostasis. The icon used to be carried in procession to its original location, the Church of Panagia Vakeriotissa, on the Tuesday after Easter Sunday. Rupert Gunnis in his 1936 publication recalled this annual event: “The icon comes from a ruined church about two miles distant, and on the Tuesday after Easter the icon is taken back in solemn procession to its original home” [21, p. 364]. Although almost a century has passed since Gunnis’s report, the tradition is still practiced by the local community (Fig. 3). It must be noted, however, that the original church building was destroyed when the Polemidia dam was built in the mid-1960s. Yet, the local tradition among the inhabitants of Palodeia was so strong that they built a replica of the church in a location unaffected by the dam, which now serves as the new destination for the annual procession. To further serve the procession’s practical needs, the icon also retained its pole (Fig. 4).¹ The icon was conserved by Kostas Gerasimou and Kyriakos Papaioakeim in 1998. However, several years prior to conservation work, it had been photographed by Sophocles Sophocleous on the 29th of June 1987. These photographs, now part of the archive of the Centre of Natural and Cultural Heritage in Nicosia, show that the hands of the Virgin Mary were covered with metal revetments (Fig. 5), thus confirming the strength of its cult and importance in the last century. Gunnis also recorded that at the beginning of the twentieth century the icon was “famous all over the countryside for the miracles it has worked […]. The icon is covered with silver eyes, for the Madonna is famed far and wide for the cures of ophthalmia which she makes.” [21, p. 364]. The silver eyes mentioned by Gunnis are preserved in the church (Fig. 6) and the icon is still considered to possess healing and miracle-making powers.²

Fig. 1

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou. Reproduced by permission of the Holy Metropolis of Limassol

Icon of Panagia Valieriotissa, Church of Saint Nicholas in Palodeia, Cyprus.

Fig. 2

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou. Reproduced by permission of the Holy Metropolis of Limassol

Icon of Panagia Valieriotissa (detail with the inset icon).

Fig. 3

Procession with the icon of Panagia Valieriotissa. Photo: Dorota Zaprzalska. Reproduced by permission of the Holy Metropolis of Limassol

Fig. 4

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou. Reproduced by permission of the Holy Metropolis of Limassol

Icon of Panagia Valieriotissa with the pole.

Fig. 5

© Centre of Natural and Cultural Heritage, Nicosia, Cyprus. Photo: Sophocles Sophocleous (29.07.1987)

Icon of Panagia Valieriotissa prior to conservation work.

Fig. 6

Votive offerings given to the icon by the faithful. Photo: Dorota Zaprzalska

Although the focus of a strong and wide cult and a particularly interesting example in terms of its form and iconography, the icon has remained virtually unknown in scholarship with the exception, of course, of the work of Sophocles Sophocleous, who included its detailed description and photographic documentation in his doctoral study [23, vol. I, pp. 256–258; vol. II, pp. 211–212; vol. III, pp. 129] as well as in his detailed work on icons from the Metropolis of Limassol [24, pp. 207–208, nos. 128–129]. He analyzed the iconography and style of the icon suggesting a 16th-century date based on similarities with the style of an anonymous Cypriot painter active in the third part of the sixteenth century, who had created a series of icons and frescos in the katholikon of the Monastery of Panagia Amasgou near the village of Monagri (1564) and in Monagri itself [25], especially with the icon of the Virgin and Christ dated 1566 from the Church of Saint George in Monagri (Fig. 7) [24, pp. 103, 206–207, no. 127]. Sophocleous also suggested that the Panagia Valieriotissa icon was repainted in the nineteenth century with only the faces of the Virgin and Christ belonging to the original painting [24, p. 207]. The rest of the original icon was either heavily cleaned or damaged, with no conclusive evidence on the matter being provided by our analysis. Sophocleous also suggested that the small icon was repainted in the nineteenth and twentieth century, but that it was a slightly earlier icon re-used and inserted into the larger one. He tentatively dated the inset icon to the sixteenth century [24, p. 208]. The present study aims to further examine this icon and its physical and material integration into the large panel with the use of technical imaging and spectroscopy.

Fig. 7

© Centre of Natural and Cultural Heritage, Nicosia, Cyprus. Photo: Sophocles Sophocleous

Icon of the Virgin Mary and Christ from the Church of Saint George in Monagri, Cyprus (121.7 × 80 × 2.7 cm).

Methods and instrumentation

The analyses described in this paper were carried out by the team of the Andreas Pittas Art Characterization Laboratories (APAC Labs) of The Cyprus Institute. The analysis was driven by art historical inquiries which aimed at deciphering the history and relation of the two panels with one another, and was based on a non-invasive and non-destructive approach integrating advanced digital imaging methods and spectroscopic methods (Fig. 8). Our approach draws from the expertise and methodological approach pursued at the Cyprus Institute and its APAC Labs, based on the central premise that the effective use of science and technology enhances research inquiries into art history and can therefore enrich our understanding of the making, technique, and materiality of works of art [26,27,28,29, pp. 113–114]. Regarding the particular icon from Palodeia, our methodology has been able to confirm and expand art historical observations, thus offering valuable conclusions in the study of this important and unique icon. Since the icon is an object of cult with great importance for the Palodeia community, most of the documentation and analyses were conducted in situ with the exclusive use of non-destructive methods. Thanks to the permission granted by the Metropolis of Limassol, the icon was transported to the APAC Labs premises in Nicosia in order to conduct additional analyses that were also limited to non-invasive methods. The investigation started with the close observation of the icon and basic digital photography to evaluate the surface condition and the state of preservation. At the next step, advanced imaging techniques were applied in order to provide preliminary suggestions of the pigments used in the icon and later interventions. Preliminary results guided the subsequent point analysis with digital microscopy and spectroscopic methods in order to identify and trace the original materials and the history of restoration and interventions which would then offer the basis for the final art-historical interpretation of results and subsequent conclusions. The aforementioned methodological approach offered a wide range of data invisible to the naked eye with regard to the painting materials, technologies and techniques used, as well as information on the subsequent use and re-use of the icon and its modifications and restorations over time. Through utilizing scientific and technological advances in the broader field of art characterization, the study of the history and provenance of paintings after their original creation has emerged as an exciting field of discovery and new knowledge within art history. We are now able to closely study the history of the preservation of paintings but also the choices related to various repainting and retouching efforts through time [30]. Such approaches are invaluable in efforts to study the layered history of composite icons.

Fig. 8

© APAC Labs / STARC, The Cyprus Institute. Image: Ropertos Georgiou

Methods used in the study.

The digital imaging techniques employed in the study of the Virgin Valieriotissa icon were exclusively non-destructive methods of examination, based on standardized methodologies and recommendations from the American Institute for Conservation (AIC).

Concisely, the methods employed were (Fig. 8):

• 1. Visible light photography (VIS) was performed on-site, by using both Medium Format (MF) Phase One 546 DF + digital back at 60 megapixels spatial resolution and Canon EOS 5D Mark full-frame CMOS digital single-lens reflex camera at 21 megapixels. The Schneider Kreuznach 80 mm f2.8 LS prime lens was used for MF and the Canon EF 50 mm f/2.5 Compact Macro for the DSLR camera body. This produced high-definition spatial resolution digital twins with maximum precision for visible colors. True color calibration was achieved under homogeneous illumination of the painting which brings about little contrast but renders the true colors. Two continuous LED array light sources "full spectrum" with white color temperature at 5500 k and Color Rendering Index (CRI) 92. Both were filtered with diffuse filters for TRUE color visible light photography. An x-rite Macbeth color passport chart was used for white balance calibration and natural realistic color rendition. The color space, which were the pixels acquired, is the Adobe RGB (1998). This particular color space has a wider gamut of colors than the standard used of sRGB color space with a wider bit depth distribution of colors. In this particular acquisition, the captured wavelength was in the range of approximately 390–700 nm (nanometers), the visible part of the spectrum. VIS images served as a reference point for the rest of the methods to be employed.

• 2. Specular Oblique illumination/Raking Light illumination (RAK) was also performed on-site. The Medium Format (MF) Phase One 546 DF + digital back at 60 megapixels spatial resolution was used for Specular Oblique and Raking Light photography (RAK). The light source was a continuous Tungsten Redhead dimmable 800W flood light, color temperature at 3200 k, Color Rendering Index (CRI) 92. An x-rite Macbeth color passport chart was used for white balance calibration and natural realistic color rendition. Raking illumination and oblique illumination create spatial shadows and interreflections among the spatial arrangement of the topography of the painting, which emphasize deviations from surface planarity by positioning a single light source at a low angle relative to the subject plane. This lighting technique was used to document the surface topography of the icon, revealing information on its condition and the techniques used.

• 3. Ultraviolet fluorescence imaging (UVF). Ultraviolet Fluorescence (UVF) is a method based on the irradiation of objects with UV light and the detection of their fluorescent radiation response in the visible, which offers the best results in completely dark surroundings. Therefore, in order to ensure the best possible conditions, this was performed in the dark room of the APAC Labs. A Canon EOS 5D Mark IV full-frame CMOS digital single-lens reflex camera at 21 megapixels and a Canon EF 50 mm f/2.5 Compact Macro lens with a Nite CC1 Filter (UV/IR cut filter) was used. We performed UVF at the visible range of the spectrum (390–700 nm) using a CC1 X-NITE filter mounted on the DSLR camera and two UV light emission units (two Madatech UV lamps with color temperature 5800 K, emission of light centered at 365 nm and color Rendering Index (CRI) 90), which irradiate UVA centered at 365 nm of the electromagnetic spectrum. Due to the fact that the UV light has a shorter wavelength than the visible light, the method is used to analyze the fluorescence response of surface materials, allowing the detection of recent restorations and the presence of UV fluorescent materials such as organic dyes, distribution of varnish, surface contamination, biological growth, and mortars that might have been used.

• 4. Infrared reflected (IR) and Infrared False Color image processing (IRFC). The infrared reflected images were acquired in situ, with a Canon EOS 6D camera with its IR filter removed, capable of capturing light at approximately the range of 390 nm ~ 1000 nm; filtered along this range with selected band-pass wavelength filters (Edmund Optics filters), ranging from 700–1000 nm, 830–1000 nm and ~ 1000 nm. The IR radiation is able to penetrate the paint layer and is generally reflected by the ground layer underneath it, while showing the underdrawings in carbon-based media that absorb IR radiation and appear dark [31]. Moreover, based on the IR images acquired in combination with the RGB channels of the VIS image (in the VIS image the green channel substituted for blue, the red channel was transferred and substituted for the blue channel, while the channel from the IR image was transferred to the red channel of the VIS image), an image processing technique was used in order to render an Infrared False Color Image (IRFC) allowing for preliminary identification of hues and the presence of illuminant metamerism. It was used for the preliminary identification of pigments that facilitated the XRF point analysis.

• 5. Infrared Reflectography (IRR) was performed at the APAC Labs premises with the use of an OSIRIS InGaAs Short Wave Infrared (SWIR) 16 megapixel panoramic camera. This imaging method of analysis is based on homogeneous irradiation of an object by using tungsten halogen light emitters, and detection of its radiation response with a suitable camera which has an InGaAs sensor, sensitive to SWIR radiation waves spanning from 900 to 1700 nm. The infrared light is characterized by a longer wavelength than the visible light and the infrared sensitivity of the IR reflected method (700 nm ~ 1000 nm), allowing it to penetrate layers of the icon deeper than the IR method. This principle, as with the IR method, allows the use of IRR for the visual analysis of preparatory sketches and underdrawings, provided that they are made with IR opaque materials (e.g. carbon-based).

• 6. Reflectance Transformation Imaging (RTI) was conducted on-site in Palodeia. RTI is a unique and intuitive visualization application of multi-light computational photography allowing the user to map dynamic light properties on the digital surface [32]. The surfaces of the large icon and the inset icon were captured under preconfigured light source locations with a fixed static Canon EOS 5D Mark full-frame CMOS digital single-lens reflex camera at 21 megapixels, and a portable electronic flash as a light source. Afterwards, images were digitally reassembled in order to create an interactive file revealing macro topography of the icon under the various lighting conditions. The software used for RTI was EOS Canon Utility for the capture and a red gloss hemispherical object (5 cm diameter) was used to record and register specular highlights from the portable flash (highlight RTI or H-RTI). Regarding the processing of the photographs, Adobe Bridge and Photoshop were used for image enhancement and color calibration, LPtracker for Polynomial Texture Maps in order to create a light source position map, and a script developed by Cultural Heritage Imaging (CHI) based on javascript in order to fit light source positions (PTMFitter) onto the image. This gave the possibility to investigate the surface for craquelure, deformation of materials and support, the brushwork technique as well as for incisions.

• 7. X-Ray Radiography (XRR) was performed at the American Medical Center in Nicosia. The system was equipped with a DX-D 40G Agfa 35 × 45 cm detector and operated at 45 kV and 56 mAh and at a distance of 1.2 m.1. XRR can penetrate the upper layers, as well as the canvas preparationand the wooden panel itself, thus revealing the structural elements of the composite icon. Due to the icon’s considerable dimensions, and mostly to its fragile wooden pole, the method was limited to capturing separate images of crucial parts, especially the area with the inset icon.

• 8. Digital microscopy was performed with a Hirox KH8700 digital microscope, equipped with a MXG-2500REZ lens (35-2500x) and a handheld MX(G)-2016Z lens (20-160x). This enabled us to study the surface condition, brushwork, layers of overpainting, and restorations, as well as the original materials – techniques of paint application, pigment particles and pigment mixtures. Since only non-invasive methods were allowed, this was limited to examining the surface and no samples were obtained for Optical Microscopy.

• 9. X-Ray Fluorescence analysis (XRF) was performed with an Elio XRF spectrometer (Bruker), equipped with a rhodium anode, operated in air at 30 kV, 100 μA, 1 mm spot size, 30 s acquisition time. The same protocol was applied to various parts of the larger icon, in order to obtain an overall characterization of the icon, and finally on the inset icon. In total, forty-nine spots were analyzed using the XRF spectrometer in order to analyze, map, and identify the pigments present in various areas of the icons. Only non-invasive methods were allowed, therefore no samples were taken and any pigments or particles were studied only on the surface of the color layers. This provided data on the icon’s original materials and the later modifications.

Results and discussion

Multispectral imaging results (IR, IRR and UVF) revealed structural information on the icon which is not visible to the naked eye. The ultra-violet image (Fig. 9) showed that the whole icon had been thoroughly covered with varnish sometime in the past century; most likely during the last conservation effort. Due to this intervention, the contrast in fluorescence measurements is far less obvious, thus not allowing the detection of retouched areas. More data were obtained from the infrared images (Fig. 10) that highlight under-drawings and changes in the painting layer compositions allowing for their precise detection. This study helped to understand which parts of the icon are original and which are later by tracing repainting and retouching. Resulting data from various methods complemented each other offering an overview of the icon’s history of painting interventions. The stylistic analysis and comparison with images taken before the icons’ conservation (Fig. 5 and 11) suggest that the inserted icon (Fig. 2 and 11) had been extensively repainted, which could be an indication of heavy use and veneration. On the large icon, no traces of the previous painting layer are visible on the IR image, supporting the hypothesis of Sophocleous that the large icon panel had been thoroughly cleaned before it was repainted. The only areas of the garments that appear to be original in the analysis are an area near Christ’s right shoulder (Fig. 12) and the maphorion of the Virgin, although its upper part bears traces of repaintings, visible on the infrared reflectogram (Fig. 10). The XRR image of this area (Fig. 13) shows the polygonal, irregular contour of the maphorion as well as the head of Christ, perhaps marking the gold under the painting layer. It should be noted that the faces themselves are recorded on the XRR image with clarity, confirming that they have been left mostly untouched. No major changes in this area were recorded by the infrared reflectogram (Fig. 10), except for a vertical line on her face that was repaired, most likely after this part cracked because this is exactly the point where two wooden panels constituting the large panel meet, as well as the part near the ear and the neck of the Virgin. This particular intervention was most likely due to a crack on this part caused by a tree ring, visible in the X-ray image (Fig. 13). This detail is also visible with the application of the RTI technique (Fig. 14) – the visible tree ring most likely caused paint loss or cracks in this particular area with clearly visible repainting.

Fig. 9

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

Icon of Panagia Valieriotissa (ultra-violet image).

Fig. 10

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

Icon of Panagia Valieriotissa (infrared image).

Fig. 11

© Centre of Natural and Cultural Heritage, Nicosia, Cyprus. Photo: Sophocles Sophocleous (29.07.1987)

Icon of Panagia Valieriotissa prior to conservation work (detail with the inset icon).

Fig. 12

© APAC Labs / STARC, The Cyprus Institute. Image: Ropertos Georgiou

Garments next to the right shoulder of Christ. From the top: Visualization of the icon detail from the RTI viewer in Visible Diffuse Gain rendering mode, in Specular Enhancement mode, and Normal map/ Depth map mode.

Fig. 13

© APAC Labs / STARC, The Cyprus Institute

Icon of Panagia Valieriotissa (detail of the faces on the large panel – XRR image).

Fig. 14

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

The area repainted over a tree ring. From the top: Visualization of the icon detail from the RTI viewer in Visible Diffuse Gain rendering mode, in Specular Enhancement mode, and Normal map/ Depth map mode.

The acquired UV, IR, IRR and IRFC images facilitated the selection of spots (Fig. 15 and 16) for the subsequent microscopy investigation and XRF analysis used for the identification of pigments in both icons. The presence of lithopone is a clear marker of overpainting, since this white pigment, which is a mixture of zinc sulphide (ZnS) and barium sulfate (BaSO₄), was in use in the second half of the nineteenth century and was first produced on a large scale in 1874 [33, p. 242]. Thus, its detection constitutes a terminus post quem for numerous areas of the large icon – the blue robe of the Virgin and the green inner part of her maphorion together with the white dots decorating its edge as well as the green-sleeved tunic of the figure of Christ and the blue band around his waist. The hands of the Virgin were painted in a mixture of lithopone and ochre (iron oxide Fe₂O₃ and clay), but the pigments used in the faces of the Virgin and Christ on the large icon were identified by XRF analysis to be of a completely different nature – a mixture of ochre, cinnabar (HgS) and lead white (2PbCO₃·Pb(OH)₂), another white pigment that was in use long before the discovery of lithopone [33, pp. 233–235; 34, p. 69]. Such a pigment mixture is consistent with the palette used by artists in the sixteenth century. The comparison of the XRF spectra taken from these two spots are shown in Fig. 17. Results support the suggestion that the faces of the two figures probably constitute an original part of the icon, while the hands of the Virgin, which had previously also been covered with metal revetments, were repainted. It cannot be excluded that this was done after the removal of the revetments, i.e. in the framework of the conservation intervention.

Fig. 15

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

XRF spots (large icon).

Fig. 16

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

XRF spots (small icon).

Fig. 17

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

Comparison of the XRF spectra (hands and faces of the large icon).

Establishing the nature of the brown color on the Virgin Mary’s maphorion seen in the large icon was more challenging because some areas appear to be painted with brown ochre and some spots reveal the presence of cinnabar, visible also in microscopic images (Fig. 18). Comparison with the icon’s archival photographs depicting the icon before its latest conservation and the infrared image (Fig. 19) indicates that some parts that now appear brown were previously covered with red pigment (the XRF analysis (point 44) identified traces of this pigment as cinnabar). Thus, some brown areas were created during the last conservation and can be clearly identified in the infrared image. The red pigment has been identified as cinnabar, a pigment also sometimes called vermilion [33, pp. 105–106, 386–387; 35]; that has been in use since antiquity and continues to be used to the present day. Therefore, it is not a marker and cannot be treated as an indication to date parts of the icon, even though it was frequently employed by post-Byzantine icon painters [36, p. 154]. Nonetheless, analysis of the style of the Virgin’s garments, with the richly bundled folds of the drapery on the shoulders, decorated rims of the textile, and sleeves, clearly indicated that this part was painted later, perhaps in the nineteenth century. The exact same pigment was used for the inscription ΜΗ(ΤΗ)Ρ Θ(Ε)ΟΥ Η ΒΑΛΗΕΡΗΟΤΗΣΣΑ and it is very likely that they were created at the same time. In this area cracks in the paint exposed the underlying layers and revealed, under the microscope, the stratigraphy of the icon – namely, a preparation layer consisting of calcium-containing white, red bole, gold and a layer of cinnabar employed for the inscriptions (Fig. 20).

Fig. 18

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

Microscopic image (maphorion of the Virgin – large icon).

Fig. 19

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

Repainted areas of the maphorion of the Virgin (VIS image, IR image).

Fig. 20

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

Microscopic image (inscription, letter A – large icon).

A visual examination of the surface of the smaller icon suggested that it had been heavily repainted, with particular attention given to the area of the faces of the depicted figures, which seemed to be heavily and badly overpainted. Unfortunately, no details of the original painting are visible, at least in the IR image (Fig. 21), which certainly raises questions as to whether the small icon’s age and cult significance could have been the main reasons for its elaborate integration into a larger panel. Microscopic imaging illustrated the complex stratigraphy of the Virgin’s maphorion which was repainted on more than one occasions (Fig. 22). Retouchings were indicated by the presence of lithopone in the flesh tones of the faces and garments. Spots consisting only of brown ochre and lead white can be considered to be from the original painting layer since these are naturally occurring pigments. The IRFC image of the inserted icon highlighted two colors on the left sleeve of Christ’s tunic thus suggesting the use of different pigments (Fig. 23). In the visible light image the whole area is blue, but with the application of the IRFC technique some parts remain blue, similarly to the frame around the icon that is green in the visible light image, but appears as blue in the IRFC image. On the other hand, some parts of the same area turned red. A number of pigments appear in IRFC as various shades of red, such as indigo, cobalt blue, phthalocyanine blue and ultramarine [37, p. 45]. Due to the detection of cobalt with XRF analysis, it can be suggested that this is cobalt blue – a pigment discovered at the end of the eighteenth century and further developed and circulated commercially by the beginning of the nineteenth century [33, p. 112–113]. XRF revealed that the same material was used in parts of Christ’s garments in the large icon which are blue in visible light but appear red in IRFC (Fig. 24), thus suggesting that they were painted with the same material. Additionally, examination by XRF confirmed that copper green (e.g. malachite) combined with lithopone was used for the green areas (a small part of the Virgin’s maphorion and the frame surrounding the inserted icon) and traced in the part of the blue tunic of Christ – the exact spot that remained blue in the IRFC.

Fig. 21

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

IR image (inset icon).

Fig. 22

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

Microscopic image (maphorion of the Virgin – inset icon).

Fig. 23

© APAC Labs / STARC, The Cyprus Institute. Image: Ropertos Georgiou

IRFC image of the inset icon.

Fig. 24

© APAC Labs / STARC, The Cyprus Institute. Image: Ropertos Georgiou

Icon of Panagia Valieriotissa (IRFC image).

The orange frame and red letters ICXC of the small icon were also examined by XRF analysis. The frame and the inscription were painted with a mixture of red ochre and lead white and restored heavily with chrome red (PbO∙PbCrO₄) and titanium white (TiO₂). The detection of titanium white in both of them strongly suggests that these elements were added in the twentieth century, since this white pigment was mass-produced for the first time at the beginning of the twentieth century [33, p. 366; 38, pp. 296–297]. It was not possible to trace the letters on the image taken before the conservation due to the bad state of preservation (Fig. 11). It was clear, however, to observe and conclude that Christ’s halo bears some changes in its execution. The comparison of the two images (Fig. 2 and 11) showed a visible distinction between them, thus permitting the suggestion that the letters IC XC were added during the conservation of the work. During this intervention, the restorers decided to considerably change the layout and representation of Christ’s halo. This particular area of the painting was carefully studied under a microscope, but only traces of the original inscription were identified (Fig. 25).

Fig. 25

© APAC Labs / STARC, The Cyprus Institute. Image: Svetlana Gasanova

Microscopic image (fragment of the inscription IC XC – inset icon).

The back of the icon was also examined (Fig. 26). Originally, three transverse battens had been fixed with nails onto the icon’s rear side, with the lower batten now missing. The back was secured at a later date with a piece of fabric on a white preparation layer containing calcium white, but also zinc white (ZnO) that was not widely used as a pigment until the end of the eighteenth century with its widespread use starting after the 1840s [33, p. 406; 39, p. 171]. We can therefore suggest that the back of the icon was decorated in the nineteenth century. The red color on the pole of the icon was identified as containing red lead (Pb₃O₄), that can be found as a mineral (minium) and its synthetic analogue was used for pigmenting purposes from ancient times [33, p. 264; 40, 41], so it is not a marker and does not provide any clear dating indication.

Fig. 26

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

Reverse of the icon of Panagia Valieriotissa.

Revealing the unseen – previous use of the inset icon and its dating

When performing the icon’s technical photographic documentation, special attention was given to the inserted small icon. The raking-light photo (Fig. 27) and the RTI show that there are no cracks between the two icons. The importance of this step of the digital documentation process is evident when images are compared to archival photographs taken before the object’s restoration (Fig. 11), where it was difficult to determine if the dark outline around the icon is the result of shading or empty space between two panels. Even though Sophocleous did suggest that the icon consisted of two panels,³ the work’s state of preservation did not allow a conclusive assessment of whether it comprised of two panels or if it was a single icon with a carved cavity imitating the form of a composite icon. In such cases, X-Ray Radiography can allow the acquisition of sub-surface information on the internal structure of an icon [42]. The XRR image (Fig. 28) of this part shows a white area around the inset icon, but it is hard to speculate on its nature – it could be a material exhibiting high X-ray absorption, a heavy metal, or a very thick layer of material used to insert and secure the small icon into the larger panel. Nevertheless, the X-Ray radiograph clearly confirms that we are dealing with two separate icons inserted one into another. The X-ray image displays the nails hammered into both icons. A nail attached to the upper part of the inset icon is a direct confirmation of the existence of the separate, re-used small icon, but details of particular interest are two metal elements arranged symmetrically on the right edge of the inset icon. These elements can be interpreted as metal parts characteristic for attaching wings to icons [42, p. 245, fig. 6] and, therefore, suggest that the inset icon had been a part of a polyptych, probably a diptych prior to the act of insertion. It must also be noted that the possibility of a triptych cannot be discarded based on indications for connections only on one side, as the small icon might have been one of the two end pieces, so this question remains open. Nevertheless, it is clear that something used to be attached to the right side of the icon. Due to its small dimensions, the inset icon was most likely intended for personal use and was originally used in private spaces, rather than in public ecclesiastical contexts. The small dimensions and private use bring to mind a group of Marian icons distinct for their mixed stylistic character, combining Italian and Palaiologan features, known in earlier scholarship under the label of the “Adriatic school” that have been revised and reinterpreted by Michele Bacci, who has provided a systematic analysis of such works [43,44,45,46,47, esp. pp. 73–96]. Icons of this type date mostly from ca. 1350 to the end of the fourteenth century. They circulated widely in the Mediterranean and, although far from being homogenous in style, share distinctive features, such as combining the Palaiologan modeling technique with the taste for rich ornaments imitating luxurious textiles and Gothicizing vestments typical of Venetian Trecento artists [44, pp. 145–146]. Unfortunately, these stylistic features cannot be observed in the case of Panagia Valieriotissa, due to the later repainting, and none of the published examples shares the exact iconography of the Palodeia icon. It is possible, however, to notice other parallel details with the icons of the above-mentioned group as well as to underline the timeframe for the proliferation of the use of such small icons for private use. Most of the icons of the aforementioned group are small rectangular panels, often parts of diptychs or triptychs meant for private or domestic piety. To the best of our knowledge, no technical analysis including X-ray radiography, that could serve as a potential comparison of the metal elements, has been conducted for any of these artworks, but it is possible to still observe in some cases metal rings, which are residual parts once serving for the attachment of wings, at the exact position of the metal elements of the Palodeia icon [45, figs 1–5]. In many cases, the Virgin Mary and Child are depicted within an inscribed arch of a very similar form to the Palodeia icon [44, fig. 2, S3, b, γ, fig. 3, b4, z, fig. 4, A, B2, C, c, fig. 5, β; 47, figs 28, 33 (= 39, 52), 41, 43, 45, 57]. It is possible that the inset of the Palodeia icon belongs to this particular group, and therefore could be dated to the second half of the fourteenth century, but we must acknowledge the limitations of non-invasive analysis, so any suggestion must be considered tentative. Nevertheless, the analytical results clearly show that the inset is a re-used icon, most likely originally intended for private use. If so, the act of insertion into a larger icon would signpost its transfer from individual devotion and private use to the public sphere of collective devotion. It bears similarity with another Cypriot composite icon known as Panagia Amirou, which is still venerated at the Monastery of Panagia Amirou, near Apsiou [48]. It consists of a 16th-century panel with an inset that most likely used to be a private icon intended for domestic use. The inset icon shares not only the already discussed characteristics of the above-mentioned group of such icons from the second half of the fourteenth century, but also the substitution of the traditional Byzantine kekryphalos with a white veil that occurs in a large number of them [44, pp. 122–126, Fig. 2; 45, pp. 170–172]. Additionally, two small metal rings arranged symmetrically on one side of the inset icon strongly suggest that a wing was once attached to it and, therefore, it was a part of a diptych, or perhaps even a triptych, prior to the act of insertion. It appears that the inset icon of Panagia Amirou used to be covered with a movable wing, or rather two wings [48]. Such wings, in the form of one or two-winged doors, appear in a large number of Cypriot composite icons and the issue of the concealment and revelation of the inset seems to play a special role in Cypriot examples [3]; however, at least at the present state of research, it is impossible to determine whether such wings were attached to the Palodeia icon as well.

Fig. 27

© APAC Labs / STARC, The Cyprus Institute. Photo: Ropertos Georgiou

Icon of Panagia Valieriotissa (detail with the inset icon – raking-light image).

Fig. 28

© APAC Labs / STARC, The Cyprus Institute

Icon of Panagia Valieriotissa (detail with the inset icon – XRR image).

Conclusions

The present paper provides an overview of the interdisciplinary study of the composite icon of the Panagia Valieriotissa from Palodeia, Cyprus. The application of a multi-technique methodological approach drew from the research work at the Andreas Pittas Art Characterization Laboratories (APAC Labs) of the Cyprus Institute, and is based on the integration of digital imaging techniques with non-invasive material analysis of the work of art. The results enhanced our knowledge of the materials and techniques used in the making of the Panagia Valieriotissa icon, but also shed light on the history of its later interventions and repainting, thus contributing to its dating and art historical study. The present work helped to establish the original parts of the icon composition as well as to map areas of later interventions, primarily through the identification of pigment materials which are of a much later date than the estimated sixteenth century date of the Palodeia icon. We can conclude that the pigments of the faces of the figures of the Virgin and Christ of the large panel are consistent with the palette of 16th-century artists, thus confirming art historical suggestions identifying them as parts of the original icon. The materials used in the overpainted areas were found to differ as they can be identified with pigments used in much later centuries. The use of lithopone in the figures’ garments confirms that most of these painting details can be attributed to the end of the nineteenth century. The icon was also retouched in the twentieth century – this was proven by the presence of pigments such as titanium white in some areas of the smaller icon. Comparison with archival photographs of the icon suggests that this was most likely done at the end of the twentieth century, perhaps during the conservation of the icon. Nonetheless, the most important result of the present study was the conclusive confirmation of the Panagia Valieriotissa icon consisting of two panels, one inserted into the other, thus securing its inclusion in the category of so-called composite icons. Moreover, the analysis revealed the unseen details of the inset icon strongly indicating that it used to be a part of a diptych, or maybe even a triptych, prior to the act of insertion and showing potential similarities to a group of 14th-century icons of a mixed Byzantine and Venetian character, thus suggesting that the inset could be much older than the embedding panel and used to be a private icon. This conclusion carries great significance as it establishes a chronological relation between the two panels, with the small icon being the older one. Moreover, this carries implications as for the choices behind the making of the larger icon, which in effect appears to be created to frame the small one. It is possible that the small icon’s age and cult value were the main reasons for the making of the composite icon. In this context, the heavy repainting of the small icon and the lack of any effort to preserve traces of its original composition remains an open question. Interestingly, repainting interventions on the large panel left most of the faces of the Virgin and Christ untouched. Unfortunately, our analyses did not yield any further information to help us better understand the choices behind the repainting of the small icon, thus leaving us no other option than to hypothesize that the original painting was simply not preserved. Future studies allowing sampling and minimally invasive methods could offer more accurate comparisons of materials, techniques, and ways of production for a larger number of composite icons, thus permitting us a better understanding of choices of artists in the reuse, repainting, or even complete removal of earlier painting layers.

In his valuable study on composite icons [1, 2], Panagiotis Vocotopoulos focused on the style and iconography of selected examples without, of course, having access to the vast wealth of information now possible with the use of advanced science and technology. The systematic exploration of the materiality of composite icons can enrich our understanding of the functional as well as symbolic aspects of these cult objects consisting of two separate icons, frequently created in different times. In our opinion, the material dimension of the physical relation between the two icons in the Panagia Valieriotissa is probably the most intriguing aspect of composite icons as it mirrors the efforts of artists and their patrons to virtually frame older icons within larger ones, thus lending complex meaning to their use in Byzantine and post-Byzantine religious life. Although there is still a lot to explore around the broader topic, the present paper, though limited to a single example, provides new perspectives in the study of composite icons as it aspires to introduce the critical application of advanced digital imaging techniques and art characterization. The proposed methodology, one that pushes the boundaries of art history research, allows for multiple levels of documentation, analysis, and interpretation in the study of icons of this type. As a result, the study aspires to serve as a pilot case study for the development of a research methodology for the study of other composite icons. Non-invasive and non-destructive methods can draw our attention to different aspects of the composite icons phenomenon and provide a better understanding of their unusual form, ways of production, and their subsequent use and re-use. Overall, the proposed multi-disciplinary methodology moves beyond traditional art historical iconographic and stylistic analysis to include spectroscopic analysis and technical photography, thus providing a holistic approach to the materiality and technique of icons, whether composite or not. In addition, this approach offers valuable information on icons’ state of preservation thereby facilitating their restoration and conservation. Moreover, the methodology applied herein shows the possibilities, as well as the limitations of non-invasive analytical methods, thus highlighting the value of the present study not only for icons but also for various cultural heritage objects.