In the early days when inferior materials were used for sculpture, colour was applied to them to conceal the poverty of the stone and to produce a more pleasing surface than that offered by the rough material at the artist’s disposal. These coarser materials were not capable of such careful finish, or of producing such a lively play of light and shade, as the marbles later used, and the only way to give them animation was by the application of colour all over the surface. It became, therefore, a regular practice for early Greek sculptors to paint their statues. When, however, they began to use more beautiful materials, such as marble, they recognised that it was a pity to conceal its texture by the extensive application of colour. They therefore adopted the practice of submitting the surface of the marble to a process of polishing, and adding colour only in parts, the effect being that the beauty of the marble is enhanced by the contrast between its polished surface and the coloured parts of the statue. The range of colours used is somewhat limited and conventional. For example, in the early pediment groups from the Acropolis, we find red used for human flesh; and the colours used in the draperies of the Acropolis female statues are limited to red and blue. Both eyes and hair are invariably red. We may infer, therefore, that colour was not added with a view to reproducing nature faithfully, but simply to decorate the statues. If, therefore, the artist felt that a white surface of marble with a few patterns sprinkled over it produced a more pleasing effect than a surface coloured all over, he would use this method of decorating his work, even if it were not realistic; and he would prefer to treat large surfaces of drapery in this way, rather than colour them all over. When, therefore, in these statues, we find that the small surface of the chiton which appears on the upper part of the figure is coloured all over, we need not conclude that the skirt belongs to another garment because it is differently ornamented; had so large a surface been painted all over, the effect would have been far less pleasing. The difference in the decoration of different parts of the same garment need in no way surprise us; it occurs very frequently in the black-figured vases, where we get purple used for the upper part of a garment and black for the lower, simply with the object of producing variety. The argument from the application of coloured ornament will not help us, then, in this case, especially when we find that it can be used to support either view. Professor Baldwin Brown has pointed out that some terra-cotta figures[132] in the Acropolis Museum, which are draped in the same style as the archaic statues, have the under-garment covering the shoulder and the skirt painted in one colour, and the part which passes round the figure under the left arm in another, and he uses this fact as a piece of evidence to show that