Although, then, the illumined space before mentioned appears indeed refracted and moved from its place, but not coloured, yet on the horizontal edges of this space we observe a coloured appearance. That here again the colour is solely owing to the displacement of a circumscribed object may require to be more fully proved.
The luminous body which here acts is circumscribed: the sun, while it shines and diffuses light, is still an insulated object. However small the opening in the lid of a camera obscura be made, still the whole image of the sun will penetrate it. The light which streams from all parts of the sun's disk, will cross itself in the smallest opening, and form the angle which corresponds with the sun's apparent diameter. On the outside we have a cone narrowing to the orifice; within, this apex spreads again, producing on an opposite surface a round image, which still increases in size in proportion to the distance of the recipient surface from the apex. This image, together with all other objects of the external landscape, appears reversed on the white surface in question in a dark room.
How little therefore we have here to do with single sun-rays, bundles or fasces of rays, cylinders of rays, pencils, or whatever else of the kind may be imagined, is strikingly evident. For the convenience of certain diagrams the sun-light may be assumed to arrive in parallel lines, but it is known that this is only a fiction; a fiction quite allowable where the difference between the assumption and the true appearance is unimportant; but we should take care not to suffer such a postulate to be equivalent to a fact, and proceed to further operations on such a fictitious basis.
Let the aperture in the window-shutter be now enlarged at pleasure, let it be made round or square, nay, let the whole shutter be opened, and let the sun shine into the room through the whole window; the space which the sun illumines will always be larger according to the angle which its diameter makes; and thus even the whole space illumined by the sun through the largest window is only the image of the sun plus the size of the opening. We shall hereafter have occasion to return to this.
If we transmit the image of the sun through convex glasses we contract it towards the focus. In this case, according to the laws before explained, a yellow border and a yellow-red edge must appear when the spectrum is thrown on white paper. But as this experiment is dazzling and inconvenient, it may be made more agreeably with the image of the full moon. On contracting this orb by means of a convex glass, the coloured edge appears in the greatest splendour; for the moon transmits a mitigated light in the first instance, and can thus the more readily produce colour which to a certain extent accompanies the subduing of light: at the same time the eye of the observer is only gently and agreeably excited.
If we transmit a luminous image through concave glasses, it is dilated. Here the image appears edged with blue.