No line and no form in the composition of the temple are exactly what they appear to be: . . . No horizontal line is really horizontal, and no vertical line really vertical, . . . every huge and massive feature is changed and almost imperceptibly deflected from the appearance it bears.[1]
The free edges, namely, of the edifice instead of being straight, as in a modern architectural design, are all curved gently in the arcs of large circles. The edges and vertical faces of the steps leading to the portico thus have a gentle convexity outward. The surface of the platform itself has the form of a very flat vault. The columns do not stand exactly vertical, but slant inward and their outlines are curved so that their actual thickness is greatest about one third the distance from the base. So great is the radius of curvature that to the casual glance there is no departure from straightness in the outlines.
Such being the objective mechanical facts, let us see what relation they may have to the visual physiology of the onlooker.
The physiological conditions may be made clear by means of a simple experiment. Let a cross formed of two strips of colored paper which intersect at right angles be fastened against a neutral tinted wall at the level of the eyes of the observer who stands at a distance of, say, ten feet. The gaze is fixed intently for some seconds upon the center of the cross. The image of the latter is thus impressed upon the retina, so that when the glance is directed elsewhere upon the wall a "negative after-image" of the cross is projected with startling distinctness upon the surface. When the eyes move so that the optic axes run along either the horizontal or vertical lines extending from the center of the cross the limbs of the latter maintain their true directions in the afterimage. But when the optic axes are directed obliquely upward or downward, the cross seems to be inclined upon the wall, the vertical limb leaning at a greater angle than the horizontal. When the orbital movement is upward and to the right, the vertical part of the cross inclines to the right, it may be as much as fifteen degrees; the horizontal limb inclines downward to the right as much as five degrees.[2] When the oblique movement of the optic axes is upward to the left the inclination of the cross is to the left. Oblique downward movements give complementary results. The amount of angular inclination of the after-image is proportional to the range of oblique movement. The physiological explanation of this phenomenon is not here important; the results are such as would occur if the eyeball in its oblique motions rotated slightly like a wheel about its visual axis.
As an observer stands before an architectural structure, his gaze roving over its lines and surfaces, the extremely complex nerve-muscle
