Those arrises were accordingly converted into _fillets_, which are not actual members, but merely spaces left between the channels or flutes themselves, which last are consequently narrower than in the Doric column; and their comparative narrowness is further increased by their being augmented in number, from that of twenty to twenty-four. Thus the change from the Doric to the Ionic column may be accounted for, rationally at least, and æsthetically, if not historically. We do not, indeed, profess to know and determine the actual origin of the volutes of the capital, and therefore leave those who put faith in Vitruvius to believe, if they can, that they were derived from the imitation of the curls in a lady’s head-dress; or, as others will have it, that the idea was borrowed either from rams’ horns, or the slender and flexile twigs of trees placed upon the capital for ornament! We also leave those who are not satisfied with our way of accounting for the base given to the Ionic column to fancy that this member was intended to imitate the ancient _chaussure_ or sandals. The Ionic capital is far more complex than that of the Doric, and not only more complex, but more irregular also: instead of showing, like the other, four equal sides, it exhibits two faces or fronts parallel to the architrave above it, and two narrower _baluster_ sides, as they are termed, beneath the architrave. Some consider this irregularity a defect, which, if such it be, is to be got over only by either turning the volutes diagonally, as in some Roman and modern examples, or by curving concavely the faces of the capital, instead of making them planes, so as to obtain four equal faces or sides, as is done in the capitals of the inner Order of the Temple of Apollo at Bassæ. At least that method, and the other one of turning the volutes diagonally, are the only methods that have been practised for giving perfect regularity to the Ionic capital by means of four equal faces; for, though difficult, it is possible to accomplish the same purpose differently, by making the abacus quite square, as in the Doric Order, and letting the volutes grow out of it on each side or face, their curvature commencing not on the upper horizontal edge, but descending from the vertical edges of the abacus. In fact, the volutes might be fancied to have originated in a prolonged abacus, first falling down on each side beneath the architrave, and then coiled up on the back and front of the column for the two faces, which thus became greater in width; after which a smaller ornamental abacus was introduced as a crowning member, immediately beneath the architrave. As it is now treated, the great extent of the two flat voluted faces prevents the capital from being square. Let us endeavour to explain this: as average measurement, we may put down 50 minutes, or 10 less than the lower diameter, for that of the upper diameter of the shaft; 65 for the sides