Refraction can visibly take place without our perceiving an appearance of colour. To whatever extent a colourless or uniformly coloured surface may be altered as to its position by refraction, no colour consequent upon refraction appears within it, provided it has no outline or boundary. We may convince ourselves of this in various ways. 196. Place a glass cube on any larger surface, and look through the glass perpendicularly or obliquely, the unbroken surface opposite the eye appears altogether raised, but no colour exhibits itself. If we look at a pure grey or blue sky or a uniformly white or coloured wall through a prism, the portion of the surface which the eye thus embraces will be altogether changed as to its position, without our therefore observing the smallest appearance of colour. XIII. CONDITIONS OF THE APPEARANCE OF COLOUR. 197. Although in the foregoing experiments we have found all unbroken surfaces, large or small, colourless, yet at the outlines or boundaries, where the surface is relieved upon a darker or lighter object, we observe a coloured appearance. 198. Outline, as well as surface, is necessary to constitute a figure or circumscribed object. We therefore express the leading fact thus: circumscribed objects must be displaced by refraction in order to the exhibition of an appearance of colour. 199. We place before us the simplest object, a light disk on a dark ground (A).[1] A displacement occurs with regard to this object, if we apparently extend its outline from the centre by magnifying it. This may be done with any convex glass, and in this case we see a blue edge (B). 200. We can, to appearance, contract the circumference of the same light disk towards the centre by diminishing the object; the edge will then appear yellow (C). This may be done with a concave glass, which, however, should not be ground thin like common eye-glasses, but must have some substance. In order, however, to make this experiment at once with the convex glass, let a smaller black disk be inserted within the light disk on a black ground. If we magnify the black disk on a white ground with a convex glass, the same result takes place as if we diminished the white disk; for we extend the black outline upon the white, and we thus perceive the yellow edge together with the blue edge (D). 201. These two appearances, the blue and yellow, exhibit themselves in and upon the white: they both assume a reddish hue, in proportion as they mingle with the black.[2] [Illustration] 202. In this short statement we have described the primordial phenomena of all appearance of colour occasioned by refraction. These undoubtedly may be repeated, varied, and rendered more striking; may be combined, complicated, confused; but, after all, may be still restored to their original simplicity. 203.