The variations in detail which appear in the different statues can easily be produced by arranging the folds in a slightly different fashion. In some cases, as for example in No. 674 (Lechat, pl. 1), the folds hang quite upright instead of obliquely, and the box-pleat appears in the middle instead of hanging from the shoulder; this can easily be produced by turning the folds first in one direction and then in the opposite. The folds of the frill sometimes hang in the opposite direction to those of the main part of the mantle; this is simply a mistake on the part of the artist. Occasionally the frill does not appear at all, for example in No. 686 (Lechat, fig. 37), but the cloak hangs straight down from the broad band. In this instance we must suppose that the overhanging mass of material has been cut away entirely before the folds were attached to the band. Sometimes the two ends were sewn together along the lines _be_ and _b′e′_, and in this case the last fastening, indicated by the letters _d_ and _d′_, approached nearer to the points _b_ and _b′_, so as to leave an opening only sufficient for the arm to pass through. The detail of the cloak which presents most variety is the little frill-like edge which falls over the band. Sometimes it appears to be a natural continuation of the vertical folds which hang down below it, and it falls over the band so as almost to hide it; sometimes it is shorter, and reveals the band and forms a sort of leaf-like pattern above it; in other cases it disappears entirely. Its most realistic representation is in one of the Victories in the Acropolis Museum, where the corners _c_ and _c′_, formed by cutting the curves, are actually indicated on the shoulder, and the frill lies in an irregular zigzag, almost exactly as it was found to fall in practice. In two cases in the Acropolis Museum at Athens, and in a statue at Delphi, the band does not pass under the arm, but from shoulder to shoulder, and the cloak covers both arms symmetrically, being fastened down both alike with a series of brooches. In these cases the box-pleat falls in the middle, and the curve must necessarily have been considerably smaller, since the upper edge lies much higher up towards the neck. When the cloak was worn in this way, it was probably sewn up down both sides, and the curves for the neck, back and front, were naturally equidistant from the two side-seams. The openings for the arms would come at the ends of the top edge, as in the case of the Ionic chiton. [Illustration: FIG. 35.—Vase-painting—British Museum. _Face page_ 93.]