8o THE PARTHENON AND ITS SCULPTURES. tions along the flanks vary among themselves, following no apparent law ; the probable reason being that their exact dimen- sions were taken from the stones intended for the epistyle, which were quarried of slightly varied lengths. The metopes of the fronts vary by two or three inches. At one of the angles the architrave projects more than four inches further on the abacus on one side than the other. The abaci of the columns are largest at the east, and smaller at the south and west. At the north and south they diminish from east to west, but not regularly. To whatever motive, says Penrose, we may attribute these irregularities, we learn that the Greek architects " did depart "^1 Fig. 66. — Lions' Heads : They are turned a little towards the Angles of the Building. somewhat from that strict regularity of proportion of which abundant examples may be found in the plan and elevation of the temple." The idea of proportion here referred to was that of making adjacent parts bear some simple ratio to each other ; thus, for example, the sides of the rectangle formed by the upper step of the stylobate were exactly as 9 to 4, and the heights and diameters of the columns, the widths of the abaci, and other parts, seem to have been designed so as to have dimensional relations each with the other. In regard to the adjustments by inclination and curvature first spoken of, it has been usual to consider them as being intended to correct certain optical defects which it is supposed
