There can be no doubt of the effect that stitching has had on the use of rosettes, but other varieties are probably independent of that. The great series of rosettes is in the moulded glazed ware of Tell el Amarna; there several dozen varieties are found, varying from four petals to thirty-two. The more elaborate of these have an unmistakable daisy centre of yellow in the midst of white petals, and this indicates what was probably the flower in mind for most of them. The rosette is found in varied use. On metal vases it is very general, and may either be a separate ornament of beaten work riveted on, like the rosettes on the silver ox head at Mykenae, or else embossed _repoussé_ in the metal. Carved in wood or ivory, rosettes decorated the furniture; and they are constantly found as centre ornaments in square patterns, and along borders with the lotus or other subject. [Illustration: 106.—L.D. II. 130.] [Illustration: 107.—P. 84.] In patterns a frequent form is only four petals, or a cruciform flower, as at Benihasan in the XIIth dynasty; and this is varied by alternations of square and diagonal arrangement. [Illustration: 108.—P. 84.] A graceful, simple form, which again recalls leather _appliquée_, is yellow on a blue ground. [Illustration: Fig. 109.] [Illustration: Fig. 110—P. 81.] [Illustration: Fig. 111.] An allied pattern is the disc surrounded by spots. This is very usual on early Greek pottery, and is found on the Aegean pottery also. This is very rarely seen in pure Egyptian design, and only in the XVIIIth dynasty, when Mykenaean influence was strongest. On Neferhotep’s ceiling two forms are found, put between the horns of the bulls’ heads, like the rosette on the Mykenaean ox head. Elsewhere it is usually seen on the scarves of the negroes as a characteristic decoration, and on the dress of the Amu (C.M. cclviii.). Hence it appears to be distinctly a foreign ornament, like the other spot pattern on a zigzag line. Only three examples are published from Egyptian decoration, and those may well be due to foreign influence. We now reach the largest and most complex growth of Egyptian ornament in the lotus, so widely spread that some have seen in it the source of all ornament. Without going so far, we shall find plenty in it to tax our reasoning and imagination. If I prefer, in dealing with this, to ignore the developments of it seen outside of Egypt as aids to understanding it, this is only because those foreign examples are so much later that they are a reflex of various Egyptian periods, and cannot show anything certainly as to the long anterior course of development in Egypt itself.