girdle clearly represented below the box-pleat by two parallel, horizontal, incised lines. On the frieze of the same building some of the figures are represented wearing the Doric peplos as an over-garment; in these cases also it is shorter than the chiton, which invariably appears below it at the feet. An archaic statue from Rhamnus, in Attica, now in the British Museum, has the crinkly chiton showing at the feet, and over it a himation with a deep overfold reaching considerably below the waist; in addition to this overfold a pleated frill appears over the breast, but no band is visible; the frill, however, is deeper than is usually the case in the Acropolis statues, and might be intended to conceal a band. This over-dress is sewn up at the side, and in that respect resembles the Doric peplos. It is significant that in this case, where the garment might with more reason be regarded as a Doric peplos let down from one shoulder, the chiton is seen appearing below it at the feet, and the over-dress does not reach to the ankles. In the few cases where the feet of the Acropolis statues are preserved, it will be noticed that the skirt is held up fairly high towards one side, so as to display the ankle. If a long under-garment were worn, we should expect its lower edge to be seen here; but in no instance is that the case, so that we may conclude that the skirt itself is the under-garment. Those who maintain that the skirt belongs to the upper garment support their opinion by the fact that very frequently the ornamentation on the two different parts is the same; the natural colour of the marble is left as a ground, and the decoration consists of coloured borders and patterns dotted somewhat sparsely over the surface. The part of the dress which appears on the left shoulder is frequently painted all over, and we might have expected that if the skirt belonged to the same garment it would also be painted all over. But before accepting this argument as conclusive, it will be well to consider the nature and purpose of polychromy as applied to Greek sculpture. [130] Perrot and Chipiez, VIII., pls. 5 and 12; Lechat, 22, 29, 30, etc. This feature comes out clearly in fig. 31. [131] Perrot and Chipiez, VIII., fig. 303; Lechat, fig. 31.