Page:Greek Buildings Represented by Fragments in the British Museum (1908).djvu/212

 196 THESEUM, ERECHTHEUM, AND OTHER WORKS. be mentioned, The columns are really complete, the closures of the small round cell being set between them, so that the columns stand in half round channels at the joints, and covering them. The architrave and frieze are in single circular blocks, and the roof is in another. It is really de- signed as a jointless fabric, that is, no joint occurs except behind the columns and at a moulding. The marble roof-stone Stuart calls a cupola, but this raises a false idea, although it is hollowed out inside for lightness. Outside it approximates to a very flat cone, but having a slight curvature, exactly the curvature of the cover of a vase or one of the turned marble toilet boxes. It fits on and into the circular cornice-stone like a lid. We know little about the develop- ment of circular buildings, but this one seems to have been designed as a pyxis, with a lid and a knop on the top. The low-sloping cone or pyramid is the classi- cal form for a round or polygonal structure. The octagonal " Tower of the Winds " was roofed with a low-pitch pyramid. The rotundas at Olympia, Epidaurus, and Samothrace must have been covered by low cones. So must the Odeon and the Skias at Athens. It is a mistake to think of the cupola, especially as seen from the outside, as a classical form. The ornament carved on the roof-stone, we are told, represents tiles, but it is rather a leaf pattern. We find it first, so far as I know, on the volute rolls of the capitals of the Nereid monument, but as a painted pattern it appeared on the Ionic capitals of the pjg. inn—pian Propylsea {R.I.B.A. Jour., vol. i.). On the top is a very handsome finial with three branches, which supported a tripod. (Fig. 200.) Traces of painted decoration have been found on this monument. Fig. 201 is the capital. Fig. 198. — Monument of Lysikrates. % J