There is just one single historical statement made by the defender. It concerned the tradition of the Fathers, who were undoubtedly speaking through the mouths of S. John Chrysostom and others supported the veneration of icons. The defence adduces no other references to the past, save citations of icons working wonders or specially honoured, in a series going back to the fifth century: the reason is that the iconoclasts demanded no historical review of the subject; both sides admitted that the icon had been accepted by the Church in extreme antiquity as a pious popular custom requiring no particular control. Still, the simplest churches either did without representations and had nothing but a cross in the apse, or had only wall-paintings and curtains with figures of the Saviour and the Apostles worked upon them, but no icons. The position was evidently different by the time when S. John Damascene wrote his three discourses defending the holy icons against those who rejected them. He had to supplement the dogmatic with the historical, or practical, side of the question. He quotes the evidence of the Fathers in favour of icons, Dionysius the Areopagite, Basil the Great, Gregory of Nysa, John Chrysostom, and ends up with cases of various specially honoured and wonder working icons in early times. It is fairly clear that it was in iconoclastic times that these specially honoured ancient icons perished. It is probable that some ancient icons of the Greek Orient have survived but are not yet known to us: of them we do know only one or two, such as the genuine Byzantine Virgin Hodegetria, carried off from Constantinople in A.D. 1204 and preserved in S. Mark’s Basilica in Venice under the name of Nicopoea, or the icon of Our Saviour in the Lateran Chapel of the Sancta Sanctorum.

      However, there are few of the truly Byzantine icons of the tenth to the fifteenth centuries of which we have knowledge. Such are in the Vatican, specifically, the icon of S.John Chrysostom on a twelfth-century reliquary of the cross from the Lateran treasure,[25] and a few small icons of the fourteenth century in the Vatican Pinacotheca, in the Pisa Gallery an icon of the Archangel Michael; in Rome the famous Hodegetria in a chapel of S. Maria Maggiore and in Bologna in a church just outside the city another miraculous icon of The Virgin, late twelfth century. The other ancient icons venerated and preserved in various churches and monasteries of Rome, Venice, Bologna, Florence, Naples, Messina, Palermo, do not belong to the true Byzantine style and are mostly Italo-Cretan work of the fifteenth to seventeenth centuries.

      It is by a rare chance that we have several Byzantine icons preserved at Novgorod: an icon of Ss. Peter and Paul in the cathedral of S. Sophia; two of the Annunciation, one in the monastery of S. Anthony the Roman, one in the church of Ss. Boris and Gleb; and one of S. George in the monastery of S. George (Yur’ev). But even in Russia the greater number of early icons are Greek and not truly Byzantine: they go back to the fourteenth and fifteenth centuries and were painted in the Greek Orient. There are some actual Byzantine icons in the State Russian Museum and they may serve as a foundation for the study of the Byzantine style.[26] Such is the remarkable icon of S. Gregory Thaumaturgus (eleventh century) inscribed with his name:[27] with its severe style it is a perfect substitute for the now whitewashed mosaic representations of bishops in the cathedral of S. Sophia at Constantinople. The faultless plastic drawing of the figure can scarcely be classed as painting, in view of the paleness of the colours and the slight indication of relief, but the perfect mastery with which the folds of the drapery are rendered by the above-described gradation or modelling with shadows, brighter planes, and highlights of varying tints of buff likewise recalls the mosaics of the Capella Palatina at Palermo. Unlike the mosaics, we find bright colour upon the sunburnt cheeks and lively flesh colour although the face is pale. This icon is clearly a real portrait, and in type remarkably like the icons of S. Gregory in S. Sophia at Kiev and his enamel icon on the Pala d’Oro in S. Mark’s, Venice.

      Equally precious is an icon of the Transfiguration which was presented to the Academy of Arts by P. I. Sevas-tiánov in the middle of the last century. Like most of the Greek or other rare specimens of his collection, he had brought it from Mount Athos. The icon, about 10 inches (25 cm.) wide, is painted on a thick oaken plank sawn out of an entablature, or rather, out of the top cornice of the iconostas of a small church or side-chapel, which it had adorned as one of a series of twelve Festivals or events of the Gospel story. They had all been painted on bright red ground, a curious peculiarity of many early icons until and including the fourteenth century. This icon, by its style, cannot be later than the tenth or possibly the beginning of the eleventh century: it is completely in the spirit of Byzantine art as restored after the iconoclastic movement. Its style is just like that of the Paris manuscripts of Gregory the Great,[28] only a certain sentiment in the type, peculiar to icon-painting, distinguishes it from the work in manuscripts.

      But the most remarkable of all examples of Byzantine icon-painting was discovered by myself at Ochrida in the church of S. Clement in 1900:[29] The icons, about 40 × 28 inches (100 × 70 cm), are evidently part of the splendid old iconostas of the thirteenth or fourteenth century moved from the cathedral when it was turned into a mosque. S. Clement’s had long been known for its antiquities but the icons were on the top row of the iconostas, covered with glass and half a century’s dust, so that it was very difficult to distinguish them. When brought down and cleaned they proved to be in almost perfect preservation, both as regards the paintings and the silver adornments on their backgrounds and frames wonderfully wrought with repoussé figures of saints and with decorative patterns. The severely majestic half-figures of Christ and of the Virgin and Child can be paralleled only by the best mosaics of the eleventh to twelfth centuries at Daphni and Palermo, and the icon of the Annunciation, adorned with the very finest cloisonné enamels of the eleventh century, is of perfect elegance. The other icons of the Virgin proved to be Serbian copies of Greco-Italian types of the Virgin and Child and belong only to the fourteenth century. We must pass over various small Byzantine icons mostly from Mount Athos. The dimensions of the bigger icons that are really Byzantine (excluding those of the fifteenth century which were produced under quite different conditions) may give us some idea of the part played by icons in Byzantine art. It is evident that Byzantine churches had their so-called ‘fixed’ icons: they were called in Russia fixed (or placed, mêstnÿya from mêsto, place) icons because being permanently fixed in the intercolumniations of the iconostas, and boarded up behind, they always remained in place.29 In cathedrals in Russia these fixed icons reached dimensions up to seven feet or so (2 m.) in height and breadth. In Byzantium the iconostas generally reached almost