awkwardness of heavy lateral members which has just been in most instances condemned. In the windows of the choir, and in most of the period, this difficulty is avoided by making the fourth order a fillet which only follows the foliation, while the three outermost are nearly in arithmetical progression of size, and the central triple shaft has of course the largest roll in front. The moulding of the Palazzo Foscari (Plate VIII, and Plate IV, fig 8) is, for so simple a group, the grandest in effect I have ever seen: it is composed of a large roll with two subordinates.
XXVIII. It is of course impossible to enter into details of instances belonging to so intricate a division of our subject, in the compass of a general essay. I can but rapidly name the chief conditions of right. Another of these is the connection of Symmetry with horizontal, and of Proportion with vertical, division. Evidently there is in symmetry a sense not merely of equality, but of balance: now a thing cannot be balanced by another on the top of it, though it may by one at the side of it. Hence, while it is not only allowable, but often necessary, to divide buildings, or parts of them, horizontally into halves, thirds, or other equal parts, all vertical divisions of this kind are utterly wrong; worst into half, next worst in the regular numbers which more betray the equality. I should have thought this almost the first principle of proportion which a young architect was taught: and yet I remember an important building, recently erected in England, in which the columns are cut in half by the projecting architraves of the central windows; and it is quite usual to see the spires of modern Gothic churches divided by a band of ornament halfway up. In all fine spires there are two bands and three parts, as at Salisbury. The ornamented portion of the tower is there cut in half, and allowably, because the spire forms the third mass to which the other two are subordinate: two stories are also equal in Giotto's
