the extreme arches a single under shaft bears two upper, proportioned as truly as the boughs of a tree; that is to say, the diameter of each upper = ⅔ of lower. There being thus the three terms of proportion gained in the lower story, the upper, while it is only divided into two main members, in order that the whole height may not be divided into an even number, has the third term added in its pinnacles. So far of the vertical division. The lateral is still more subtle. There are seven arches in the lower story; and, calling the central arch a, and counting to the extremity, they diminish in the alternate order, a, c, b, d. The upper story has five arches, and two added pinnacles; and these diminish in regular order, the central being the largest, and the outermost the least. Hence, while one proportion ascends, another descends, like parts in music; and yet the pyramidal form is secured for the whole, and, which was another great point of attention, none of the shafts of the upper arches stand over these of the lower.
XVI. It might have been thought that, by this plan, enough variety had been secured, but the builder was not satisfied even thus: for—and this is the point bearing on the present part of our subject—always calling the central arch a, and the lateral ones b and c in succession, the northern b and c are considerably wider than southern b and c, but the southern d is as much wider than the northern d, and lower beneath its cornice besides; and, more than this, I hardly believe that one of the effectively symmetrical members of the façade is actually symmetrical with any other. I regret that I cannot state the actual measures. I gave up the taking them upon the spot, owing to their excessive complexity, and the embarrassment caused by the yielding and subsidence of the arches.
Do not let it be supposed that I imagine the Byzantine workmen to have had these various principles in their minds as they built. I believe they built altogether
