The processes of mosaic production involve all the same stages as wall-paintings and some extra ones. Mosaicists no doubt worked in both media, and this is made clear from the stylistic similarities of the wall-paintings and the mosaics of the Kariye Camii produced between 1315 and 1321 when the church was completed (Underwood 1967:172). Very likely the same artists also produced some of the icons in the churches they had decorated.
Until observations were made, in particular by the restorer Ernest Hawkins, during the restorations of the mosaics of Hagia Sophia at Constantinople which started in 1932 (Cormack 1981), the predominant view in the literature was that Byzantine and Italian mosaics were made in the studio, while reversed and attached to cartoons, and were then carried to the church and set in position on the walls (Underwood 1967, vol. 1:172-3,179). What disproved this was the observation that tesserae were individually set in situ on the walls to catch the light or to adjust to individual structural features in the building. One clear case of pragmatic mosaic work is in the figure of St lohn the Baptist in the fourteenth-century bema mosaic at the Fetiye Camii. This figure is in the south lunette and the artist has not filled in the flesh of the hands or the feet with tesserae but only with painted fresco (Underwood 1967, vol. 1:179; Belting and others 1978: fig 24a). Not only were these flesh parts set last but the artist has economized in the use of tesserae in a position which would have been very hard for the Byzantine viewer ever to see, as the bright light from the south window would have made these ad hoc economies features virtually invisible to the human eye. The same shortcut of using fresco instead of mosaic was made in this church in the saints in the prothesis and diakonikon where some other figures were equally difficult for the viewer to see (St Clement, St Cyril, and St Athanasios).
The mosaicist generally covered the bare walls of the church with not two but three plaster layers, although some cases of two layers only have been claimed in the modern literature. The surface of the first two layers was treated similarly with a series of impressed indentations. The purpose was to ensure that the final layer adhered firmly to the wall or the vault. However, in order to give extra adhesion of the plaster to the vaults, it was normal to hammer iron nails with large flat heads into the masonry, and the plaster was set around these nails. This was unfortunately only a short-term solution and over time the nails were liable to rust and gradually to expand through corrosion, causing the plaster to splinter and crumble and the tesserae to fall. The third and final layer, the setting bed, consisted of lime and a filler, usually marble dust, and the straw was omitted. This procedure meant that the combined thickness of the three plaster layers normally measured from 4 to 5 cm.
Where tesserae have fallen from the setting bed, it is possible to see the nature of the preliminary work carried out during production. Hence it is clear that it was the practice to apply a full painted composition to the wet setting bed before the tesserae were inserted one by one. This under-painting was to act as a guide both to the placing and forms of the figures and backgrounds and also to the colours to be used. As for the gold background to be found in many mosaics, this was indicated usually with an earth red ground, although sometimes yellow ochre was used instead. These coloured grounds meant that an extra depth was given to the values of the gold grounds of mosaics as these painted colours may have been visible through the interstices of the tesserae and would have enhanced the gold. In cases where the gold tesserae were tilted to reflect more light and at the same time fewer were used for reasons of economy, the coloured setting bed compensated for the lack of tesserae (as in the case of the narthex mosaic in Hagia Sophia at Constantinople).
As in the case of wall-paintings, it is hard to unravel from observation of the surface of the mosaic, or even of exposed setting beds, how much work was achievable in the course of one day’s work. In the Kariye Camii in the Presentation of the Virgin in the Temple, a scene which included fourteen figures as well as architectural elements and trees and which covers the whole of the vault of the third bay of the inner narthex, one easily visible joint in the gold background runs across the centre of the composition. This was taken to indicate that the scene was executed in two major sections (Underwood 1967: vol. 1:178; vol. 2: fig. 119). But it is unclear whether in fact each of these sectors was divided into smaller sections, with, for example, each figure representing a separate piece of work and with the suture in the gold background representing no more than the final operation in the mosaic when the gold tesserae were set between all the figures and the architecture. The interpretation is important, as it might be an indicator of the number of people involved in setting the mosaics—either one man and a few assistants or a large workshop operation. Both views are to be found in the art historical literature. It is the case that ornamental borders around scenes are done in separate operations after the scene is finished (Underwood 1967: vol. 1:178).
The size of the tesserae differs, depending on their position—those for the face and hair are usually the smallest and those for the gold background usually the largest. The materials for tesserae are various, but the predominant materials are glass and stone or marble. In addition brick, glazed pottery, and mother of pearl have been recorded. Manufacture of the glass tesserae was done by producing plates, which were cut into individual cubes. The basic colours were blue, green, violet, red, yellow, and black. In the areas of gold ground, sometimes there is an admixture of silver tesserae to change the appearance of the gold and sometimes a few tesserae are reversed and also act to change the appearance of the surface.
While the use of mosaic as a rich and glimmering cover for the walls and vaults of churches was perfected in Byzantium, the medium itself has a long history in Antiquity as does the manufacture of glass. By the period of the Roman Empire, both floor and wall mosaics had been developed as a popular medium. Although stone was the normal substance used in Roman floors, at both Pompeii and Herculaneum other materials are found in wall mosaics, particularly in those which embellished fountains and water features, where glass and shells are to be found. During the Roman period, there were numerous glass factories, especially in the Levant, Egypt, and Italy (James 2006:34). These produced and exported either raw glass or finished glass objects. It may be that the raw glass for Byzantine mosaics was primarily made in the Levant and imported to Constantinople in raw form. Small-scale workshops with a simple furnace using imported raw glass may have been the pattern of glass-working in Constantinople and throughout the empire (James 2006: 38). Byzantine glass, at least from the ninth century, was of the soda-lime-silica variety with high levels of magnesium. This distinguishes it from Roman glass, which had little potassium and magnesium.
For the production of tesserae the glass had to be coloured, and this implies a specialist process for manufacturing batches in different colours as well as tesserae with an inserted layer of gold or silver leaf below a thin wafer of glass. This work may perhaps have been done in furnaces on the site of the church being decorated, although the reference to the mosaicists of the churches of Kiev in the eleventh century bringing tesserae from Constantinople implies that at least some tesserae might have been transported with itinerant artists (Mango 1972:221-2). To make the tesserae on site implies an expertise on the part of the mosaicists in the making of coloured glass, for example knowing how to add copper for blues and greens, cobalt for dark blues and how to control the oxygen content of the glass melt for dark reds. The four colouring elements, iron, copper, manganese, and cobalt and their combinations in different proportions and the addition of opacifying agents such as antimony and tin allowed for the production of most colours (James 2006:39-42). But some particular colours, such as vermilion, seemed to have caused problems of mass production and were in short supply. In Carolingian Rome there is evidence of the collection and reuse of old Roman tesserae, and such tesserae were also used in Byzantine enamels. But so far scientific analysis of Byzantine tesserae suggests that the tesserae were newly made for each operation.
When the tesserae had been laid, a final procedure was to stand back and to assess the appearance of the mosaic. At this stage, the master mosaicist used pigments and a brush to touch up the appearance of any figures that did not ‘look right’. This explains the reason for painting over the tesserae to enhance the red lips of the Virgin (as in the apse mosaic of Hagia Sophia at Constantinople) and to paint in the red vermilion on the white buskins of the emperor in the narthex mosaic of the same church as well as on the cushion of Christ’s throne. In the case of the Kariye Camii and Fethiye Camii mosaics significant areas of figures and other elements were enhanced with paint (Underwood 1967:182-3).
According to Giorgio Vasari, the fifteenth-century artist Domenico Ghirlandaio used to say ‘the true painting for eternity was mosaic’, but in reality most mosaics which have survived have gone through many restoration processes, and some, particularly in Italy, have been removed from the wall and reset in modern times. In the Byzantine period, too, some mosaics were subjected to change and reworking. The clearest case is in the sanctuary mosaics of the church of the Koimesis at Nicaea, now known only through photographs and drawings made before their destruction in 1922 (Underwood 1959). It is clear here that a mosaic could be changed by two procedures: one was to cut out tesserae one by one and replace them in the different colours. For example, a cross or an inscription could be eliminated by replacing the tesserae with new ones of the same colour as the background. The feature then disappears, although a ghostly trace of the original may remain, as do the two horizontal arms of a cross inserted in the Nicaea apse by some iconoclasts which were later removed tesserae by tesserae by iconophiles. The other procedure was to cut out a patch of the mosaic and then to fill in the empty space with a new setting bed and new tesserae. This happened twice at Nicaea: the iconoclasts removed the original standing Virgin (dating perhaps around 700) and inserted a cross in the apse. This cross was removed in the ninth century after the end of iconoclasm, in part by replacing the tesserae of the horizontal arms but at the same time the vertical post of the cross was cut out and a standing Virgin and Child was inserted on a new setting bed. The evidence of these alterations and so of the three phases of work in these mosaics was still visible on the early twentieth-century photographs. Knowledge of these procedures makes it still possible to see that the current Virgin and Child in the apse of Hagia Sophia at Thessalonike replaced an
Fig. 1 The apse of the Church of the Koimesis, Nicaea
Original decoration of a cross and that in the room over the south-west ramp of Hagia Sophia at Constantinople the iconoclasts removed a sixth-century cycle of saints in medallions and replaced them with crosses in the 768/9 recorded alterations (Cormack 1981). The more puzzling example of alterations is the replacement of the faces of Christ and Zoe as well as Constantine XI Monomachos when the imperial panel in the south gallery of Hagia Sophia was updated to record a new donation to the church by the third husband of Zoe between 1042 and 1055.
Although the vast majority of mosaic decorations were executed in Constantinople, elsewhere in the Mediterranean artists travelled to carry out major programmes of work—as for example in Italy at Venice, Rome, and Sicily, and in Spain in the Great Mosque at Cordova, and also in the Middle East in Georgia and at Damascus and Jerusalem. Kiev has already been mentioned. The obvious questions to ask are how many artists travelled as a group and what they took with them. At present we have no evidence to pronounce on the size of wall-painting or mosaic workshops and this is, as already mentioned, a controversial issue. As for equipment, they needed to acquire the necessary lime, pigments, and glass and stone tesserae at each church. In addition to their working tools, recent literature has given much emphasis to the use of the compass, both for making haloes and for helping to decide the proportions of figures, perhaps based on the module of the nose (Torp 1984; Winfield and Winfield 2003). Another question is how far artists relied on model books as reference records for their figures and compositions or on memory. Deciding between these is highly controversial. It is doubtful, for example, whether before the wider use of paper in the Late Byzantine period model books could have existed in any numbers. But it is a question of some significance for on it hangs the issue of how far the Church might have controlled the contents of monumental art and how far artists controlled the styles and means of expression. What is clear is that for most of the history of Byzantine church decoration the subjects were relatively limited, with particular emphasis given to the main festival scenes of the liturgical year and representations of Christ, the Virgin, and the saints, but with some interest in giving emphasis to the local saints of the region in which the church was located. By the Late Byzantine period, cycles were expanded with new scenes in the life of the Virgin and saints and the development of the iconography of Christ himself.
References
Belting, H., Mango, C., and Mouriki, D. 1978. The Mosaics and Frescoes of St. Mary Pammakaristos (Fethiye Camii) at Istanbul (Washington, DC).
Borsook, E. 1981. The Mural Painters of Tuscany (London, 2nd edn.).
Cormack, R. 1981. ‘Interpreting the mosaics of S. Sophia at Istanbul’, Art History 4:131-49.
Demus, O. 1948. Byzantine Mosaic Decoration: Aspects of Monumental Art in Byzantium (London).
-1950. The Mosaics of Norman Sicily (London).
Dodwell, C. R. (ed.) 1961. Theophilus De Diversis Artibus (London).
Gerstel, S. E. J. 1999. Beholding the Sacred Mysteries: Programs of the Byzantine Sanctuary (Seattle-London).
Herringham, C. J. (ed.) 1899. The Book of the Art of Cennino Cennini (London).
Hetherington, R 1974. The 'Painter's manual' of Dionysius ofFourna: An English Translation [from the Greek] with commentary of cod. gr. yo8 in the Saltykov-Shchedrin State Public Library, Leningrad (London).
James, L. 2006. ‘Byzantine glass mosaic tesserae: some material considerations’, BMGS 30: 29-47.
Mango, C. 1962. Materials for the Study of the Mosaics of St. Sophia at Istanbul (Washington, DC).
-1972. The Art of the Byzantine Empire 312-1453: Sources and Documents (Englewood
Cliffs).
Merrifield, M. R (ed.) 1849. Medieval and Renaissance Treatises on the Arts of Painting (London; repub. 1999).
Meyendorff, P. (ed. and trans.) 1984. St Germanus of Constantinople, On the Divine Liturgy (Crestwood, NY).
Mouriki, D. 1985. The Mosaics ofNea Moni on Chios (Athens).
Moutafov, E. S. 2001. Europeanisation on Paper: Treatises on Painting in Greek during the First Half of the 18th Century (Sofia).
Nordhagen, P. j. 1965. ‘The mosaics of John VII (705-707 A. D.). The mosaic fragments and their technique’, Acta ad archaeologiam et atrium historiam pertinentia 2:121-66.
L’Orange, H. R, and Nordhagen, R J. 1966. Mosaics from Antiquity to the Middle Ages (London).
Papadopoulos-Kerameus, a. 1909. Manuel d'iconographie chretienne (St Petersbourg).
Safran, L. (ed.) 1998. Heaven on Earth: Art and the Church in Byzantium (University Park, Pa.).
Thompson, D. V. 1936. The Materials of Medieval Painting (London).
Torp, H. 1984. The Integrating System of Proportion in Byzantine Art (Rome).
Underwood, P. A. 1959. ‘The evidence of restorations in the sanctuary mosaics of the Church of the Dormition at Nicaea’, DOP13: 235-44.
-1967. The Kariye Djami, vol. 1: Historical Introduction and Description of the Mosaics
And Frescoes; vol. 2: The Mosaics; vol. 3: The Frescoes (London).
Intellectual Background (London).
Winfield, D. C. 1968. ‘Middle and later Byzantine wall painting methods: a comparative study’, DOP 22: 61-139.
-and Winfield, J. 2003. The Church of the Panaghia tou Arakos at Laghoudera, Cyprus:
The Paintings and their Painterly Significance (Washington, DC).
Suggested Reading
The church of the centre of Byzantium which was decorated with mosaics from the sixth century onwards and presumably seen by more people than any other monument is Hagia Sophia at Constantinople, and the bibliography on its decoration sets out all the key issues
- (ed.) 1975. The Kariye Djami, vol. 4: Studies in the Art of the Kariye Djami and its
For study (see Cormack 1981 for references). An exemplary coverage of the mosaic materials recorded in the church is given by Mango 1962. Equally important for late Byzantium is the decoration of the Kariye Camii, and this is well covered by Underwood 1967,1975. For a clarification of the nature of mosaic styles and treatment of iconography Demus 1950 remains a classic, but while his interpretative book of 1948 about the planning and evolution of church programmes is insightful and provocative, it is a very problematic account. There are many good studies of individual churches and their decoration, and interest has shifted recently to patterns of decoration and how far these are specific to the dedication and site of the church and how far to generic principles of planning (see Safran 1998 and Gerstel 1999). ,
II.8. PRODUCTION, MANUFACTURE, AND TECHNOLOGY
World History
