If, when the sky is grey, we approach a window, so that the dark cross of the window-bars be relieved on the sky; if after fixing the eyes on the horizontal bar we bend the head a little forward; on half closing the eyes as we look up, we shall presently perceive a bright yellow-red border under the bar, and a bright light-blue one above it. The duller and more monotonous the grey of the sky, the more dusky the room, and, consequently, the more previously unexcited the eye, the livelier the appearance will be; but it may be seen by an attentive observer even in bright daylight.
If we move the head backwards while half closing the eyes, so that the horizontal bar be seen below, the phenomenon will appear reversed. The upper edge will appear yellow, the under edge blue.
Such observations are best made in a dark room. If white paper is spread before the opening where the solar microscope is commonly fastened, the lower edge of the circle will appear blue, the upper yellow, even while the eyes are quite open, or only by half-closing them so far that a halo no longer appears round the white. If the head is moved backwards the colours are reversed.
These phenomena seem to prove that the humours of the eye are in fact only really achromatic in the centre where vision takes place, but that towards the circumference, and in unusual motions of the eyes, as in looking horizontally when the head is bent backwards or forwards, a chromatic tendency remains, especially when distinctly relieved objects are thus looked at. Hence such phenomena may be considered as allied to the dioptrical colours of the second class.
Similar colours appear if we look on black and white objects, through a pin-hole in a card. Instead of a white object we may take the minute light aperture in the tin plate of a camera obscura, as prepared for paroptical experiments.