As the compensatory colours easily appear, where they do not exist in nature, near and after the original opposite ones, so they are rendered more intense where they happen to mix with a similar real hue. In a court which was paved with grey limestone flags, between which grass had grown, the grass appeared of an extremely beautiful green when the evening clouds threw a scarcely perceptible reddish light on the pavement. In an opposite case we find, in walking through meadows, where we see scarcely anything but green, the stems of trees and the roads often gleam with a reddish hue. This tone is not uncommon in the works of landscape painters, especially those who practice in water-colours: they probably see it in nature, and thus, unconsciously imitating it, their colouring is criticised as unnatural. 60. These phenomena are of the greatest importance, since they direct our attention to the laws of vision, and are a necessary preparation for future observations on colours. They show that the eye especially demands completeness, and seeks to eke out the colorific circle in itself. The purple or violet colour suggested by yellow contains red and blue; orange, which responds to blue, is composed of yellow and red; green, uniting blue and yellow, demands red; and so through all gradations of the most complicated combinations. That we are compelled in this case to assume three leading colours has been already remarked by other observers. 61. When in this completeness the elements of which it is composed are still appreciable by the eye, the result is justly called harmony. We shall subsequently endeavour to show how the theory of the harmony of colours may be deduced from these phenomena, and how, simply through these qualities, colours may be capable of being applied to æsthetic purposes. This will be shown when we have gone through the whole circle of our observations, returning to the point from which we started. [1] Plate 1, fig. 3. VI. COLOURED SHADOWS. 62. Before, however, we proceed further, we have yet to observe some very remarkable cases of the vivacity with which the suggested colours appear in the neighbourhood of others: we allude to coloured shadows. To arrive at these we first turn our attention to shadows that are colourless or negative. 63. A shadow cast by the sun, in its full brightness, on a white surface, gives us no impression of colour; it appears black, or, if a contrary light (here assumed to differ only in degree) can act upon it, it is only weaker, half-lighted, grey. 64. Two conditions are necessary for the existence of coloured shadows: first, that the principal light tinge the white surface with some hue; secondly, that a contrary light illumine to a certain extent the cast shadow. 65.