[Illustration: 138.—T.A. 388.] This view is further complicated by showing not only some of the four outer sepals, but also some of the petals, usually three. Here the near sepal is shown rising in front, and then above these everted sepals are three of the inner petals of the flower. These might be increased to five or seven, but were generally an odd number; and they were at last evolved to a fan of petals, in which the treatment of the dish of fruit just shown is exactly reproduced, a side view of the flower being crowned by a top view of it showing the radiating petals in the interior. [Illustration: 139.—T.A. 375.] [Illustration: 140.—T.A. 374.] So far we are on clear ground. Now we come to a more complex form, which has also not yet been explained. In the XVIIIth dynasty (from which we must mainly draw, as we have the long series of varieties in the glazed ornaments of Tell el Amarna) a strange form appears, with reversed curling arms above the calyx. Now we have seen that a third sepal is shown from the back of the flower, and the fourth is omitted which lay in front. But this was an imperfect flower, and so a diagonal point of view was taken, in which two sepals lay nearest and were seen in side view, and the two behind them were seen over them. Sometimes they are curled alike, but more generally they are curled different ways, the nearer ones downwards, the further ones upwards. Hence we get this very mechanical form, which was greatly developed in Assyrian and Greek types of the pattern. If it can be proved that the Assyrian tree pattern is earlier than this development, we could then grant what seems a likely influence on the development of this pattern. It was so far removed from a natural view that it soon became greatly varied and amplified, as on a bracelet in the Louvre. [Illustration: 141.—P. 113.] [Illustration: 142.—P. and C. Ass. 127.] [Illustration: 143.—Tanis II. xxxi.] [Illustration: 144.—Goodyear. 75.] In Assyria this became a staple design, in which the top was greatly increased at the expense of the lotus sepals below; but still the four sepals, two front and two back, are shown. In the Greek designs, however barbarous they may seem in comparison, owing to their hopeless divergence from any rational type, yet the same elements remain, and the four sepals can be traced below the view of the petals in the flower. Thus the anthemion with its double curves is fully accounted for, the lower and upper sepals being still distinguishable in the two spirals on each side at the base of it. The later changes of this necessarily belong to Greek art, and we cannot here follow them out. [Illustration: Fig. 145.] A late development of the lotus in Ptolemaic Egypt was with a central spike through the face of petals. As this spike rises from the base, it appears to be the front sepal rising before the petals. [Illustration: 146.—P. 111.]