For, hitherto, light has been considered as a kind of abstract principle, existing and acting independently; to a certain extent self-modified, and on the slightest cause, producing colours out of itself. To divert the votaries of physical science from this mode of viewing the subject; to make them attentive to the fact, that in prismatic and other appearances we have not to do with light as an uncircumscribed and modifying principle, but as circumscribed and modified; that we have to do with a luminous image; with images or circumscribed objects generally, whether light or dark: this was the purpose we had in view, and such is the problem to be solved. 362. All that takes place in dioptrical cases,--especially those of the second class which are connected with the phenomena of refraction,--is now sufficiently familiar to us, and will serve as an introduction to what follows. 363. Catoptrical appearances remind us of the physiological phenomena, but as we ascribe a more objective character to the former, we thought ourselves justified in classing them with the physical examples. It is of importance, however, to remember that here again it is not light, in an abstract sense, but a luminous image that we have to consider. 364. In proceeding onwards to the paroptrical class, the reader, if duly acquainted with the foregoing facts, will be pleased to find himself once more in the region of circumscribed forms. The shadows of bodies, especially, as secondary images, so exactly accompanying the object, will serve greatly to elucidate analogous appearances. 365. We will not, however, anticipate these statements, but proceed as heretofore in what we consider the regular course. XXXI. CATOPTRICAL COLOURS. 366. Catoptrical colours are such as appear in consequence of a mirror-like reflection. We assume, in the first place, that the light itself as well as the surface from which it is reflected, is perfectly colourless. In this sense the appearances in question come under the head of physical colours. They arise in consequence of reflection, as we found the dioptrical colours of the second class appear by means of refraction. Without further general definitions, we turn our attention at once to particular cases, and to the conditions which are essential to the exhibition of these phenomena. 367.