“In the =Parthenon=, for instance,” (the quotation is from the writings of Professor William H. Goodyear) “surfaces or members which are set true to perpendicular are most exceptional. Perhaps the end walls are the only exception. All the columns lean inward about three inches in thirty feet toward the centre of the building. The side walls lean inward. The antæ, or flat pilasters at the angles of the ends of the walls, lean forward one unit in eighty-two units. The faces of the architrave and frieze lean backward, whereas the acroteria, the face of the cornice and the face of the fillet between architrave and frieze lean forward. Furthermore, the columns and capitals of the Parthenon are of unequal size, and the widths of the metopes and the intercolumnar spacings are also unequal.” The discovery of these variations was pooh-poohed by architects who had been trained to believe that “correct” architecture depended upon geometrical regularity and mathematical accuracy. They dismissed them lightly as “mason’s errors.” But this will not hold for three reasons. Firstly, these asymmetries only occur in the finest examples, where the design and the details are of superior refinement and the skill of the masons most unmistakable. Secondly, the number of variations increases _pro rata_ with the superiority of the design, reaching their maximum in the =Parthenon=. And, thirdly, in cases which are unquestionably due to mason’s errors the amount of the variation is practically negligible. Is it likely, for example, that the masons who brought the two ends of the Parthenon within one quarter of an inch of being exactly equal in width, would have been so careless as to let the presumably horizontal lines curve up four inches on the sides of the buildings and two inches at its ends? Or, again, would they have committed so flagrant an error as giving the stylobate a convex curve upward, since it necessitated a corresponding curve to the base of each column, a most difficult and delicate operation of cutting? The perfect adjustment of these two curves, by the way, is one of many arguments against the theory that these variations were caused by settlements in the foundations or, in the case of the Parthenon, by the explosion which wrecked it in 1687, when it was being used by the Turks as a powder magazine.