This experiment may be repeated and varied before a smooth wall, with rods of different thicknesses, and again with balls; we shall always find that the farther the object is removed from the surface of the wall, the more the weak double shadow spreads, and the more the forcible main shadow diminishes, till at last the main shadow appears quite effaced, and even the double shadows become so faint, that they almost disappear; at a still greater distance they are, in fact, imperceptible. 397. That this is caused by the cross-action of the light we may easily convince ourselves; for the shadow of a pointed object plainly exhibits two points. We must thus never lose sight of the fact that in this case the whole sun-image acts, produces shadows, changes them to double shadows, and finally obliterates them. 398. Instead of solid bodies let us now take openings cut of various given sizes next each other, and let the sun shine through them on a plane surface at some little distance; we shall find that the bright image produced by the sun on the surface, is larger than the opening; this is because one edge of the sun shines towards the opposite edge of the opening, while the other edge of the disk is excluded on that side. Hence the bright image is more weakly lighted towards the edges. 399. If we take square openings of any size we please, we shall find that the bright image on a surface nine feet from the opening, is on every side about an inch larger than the opening; thus nearly corresponding with the angle of the apparent diameter of the sun. 400. That the brightness should gradually diminish towards the edges of the image is quite natural, for at last only a minimum of the light can act cross-wise from the sun's circumference through the edge of the aperture. 401. Thus we here again see how much reason we have in actual observation to guard against the assumption of parallel rays, bundles and fasces of rays, and the like hypothetical notions. 402. We might rather consider the splendour of the sun, or of any light, as an infinite specular multiplication of the circumscribed luminous image, whence it may be explained that all square openings through which the sun shines, at certain distances, according as the apertures are greater or smaller, must give a round image of light. 403. The above experiments may be repeated through openings of various shapes and sizes, and the same effect will always take place at proportionate distances. In all these cases, however, we may still observe that in a full light and while the sun merely shines past an edge, no colour is apparent. 404.