M. Hittorff has in his possession a fragment of a figure from Selinus, retaining a flesh colour very similar to that which we have employed.

Although colour has been found on the hair, eyes, lips, and drapery of Greek fragments of marble, no traces have as yet been found on the nude portions. And those who believe that the marble of the Greeks was only stained and not painted, build up a triumphant argument on this. The explanation, however, is very simple; it is evident that the smooth portions of a coloured object would lose their colour first under the influence of time, and, in fact, all traces of colour that ever are found, are found in the folds and crevices, from which it is fairly argued that the surface of which they formed a part was of that colour.

Even in the Alhambra, which was entirely covered with colour, and which is so many centuries nearer our time than the Greek temples, colour is but rarely found on the surface: it is only by what is found in the depths and hollows, that we know how the whole was coloured.

On the terra-cottas of the Louvre there are figures where the white ground with which the whole surface of the terra-cottas was covered, remains perfect over the whole of the figures, at the same time that a fragment of flesh tint still remains upon some portion of it. Were this absent, it might equally well be argued, that the Greeks were in the habit of painting the flesh white on their terra-cottas.

HORSES.

In seeking a colour for the horses, I felt the choice lay between red, white, black, or grey; further, that whatever colour was employed, it would be in such a way as best to define and distinguish the various portions of the groups. I do not think that a single colour, or shades of the same colour, would have fulfilled this condition. White horses would have been too prominent, black too sombre. The red I have employed appeared to be the best colour for the principal horses, as best balancing by their masses the blue background, whilst the relief between horse and horse could be harmoniously obtained by the employment of grey for the back horses. Authority for this mode of treatment exists on the Greek vases and in the Etruscan tombs, where, when one horse passes before another, there is a change of colour. As the horses in this frieze are in ranks of nine, it is most probable that there was still more variety of colour than I have attempted, to keep the various groups together.

THE DRAPERIES.

I was led to adopt this mode of treating the draperies from the inspection of the Louvre collection of terra-cottas, where the draperies are very well preserved. They are mostly pale blue and pale pink, the pale blue with a pink border and the pink drapery with a blue border. I have arranged the draperies in the way I felt most conducive to the general effect, so as to bring the whole into harmony. The colours of the other portions of the dresses are suggested by the materials which they may be presumed to represent.

In placing this experiment before the public, I am quite aware how vain would be the hope that I had produced a result worthy of the Greeks; where there is so little to guide, success is well nigh impossible. The most that I could hope to attain was to produce a result that might have existed, and that would not have been discordant with the other portions of a Greek monument. My failures even would answer a useful purpose, if they served to direct other minds to work out this most interesting problem, and to induce further researches on the monuments of Greece, which have hardly yet been examined in this direction, because they have not as yet been examined with faith, but rather with reluctance.

The experiment cannot be fairly tried till tried on marble, and in conditions of space, atmosphere, &c., similar to those under which the originals were placed.