Every treatise on the harmonious combination of colours contains the diagram of the chromatic circle more or less elaborately constructed. These diagrams, if intended to exhibit the contrasts produced by the action and re-action of the retina, have one common defect. The opposite colours are made equal in intensity; whereas the complemental colour pictured on the retina is always less vivid, and always darker or lighter than the original colour. This variety undoubtedly accords more with harmonious effects in painting. The opposition of two pure hues of equal intensity, differing only in the abstract quality of colour, would immediately be pronounced crude and inharmonious. It would not, however, be strictly correct to say that such a contrast is too violent; on the contrary, it appears the contrast is not carried far enough, for though differing in colour, the two hues may be exactly similar in purity and intensity. Complete contrast, on the other hand, supposes dissimilarity in all respects. In addition to the mere difference of hue, the eye, it seems, requires difference in the lightness or darkness of the hue. The spectrum of a colour relieved as a dark on a light ground, is a light colour on a dark ground, and _vice versâ_. Thus, if we look at a bright red wafer on the whitest surface, the complemental image will be still lighter than the white surface; if the same wafer is placed on a black surface, the complemental image will be still darker. The colour of both these spectra may be called greenish, but it is evident that a colour must be scarcely appreciable as such, if it is lighter than white and darker than black. It is, however, to be remarked, that the white surface round the light greenish image seems tinged with a reddish hue, and the black surface round the dark image becomes slightly illuminated with the same colour, thus in both cases assisting to render the image apparent (58). The difficulty or impossibility of describing degrees of colour in words, has also had a tendency to mislead, by conveying the idea of more positive hues than the physiological contrast warrants. Thus, supposing scarlet to be relieved as a dark, the complemental colour is so light in degree and so faint in colour, that it should be called a pearly grey; whereas the theorists, looking at the quality of colour abstractedly, would call it a green-blue, and the diagram would falsely present such a hue equal in intensity to scarlet, or as nearly equal as possible. Even the difference of mass which good taste requires may be suggested by the physiological phenomena, for unless the complemental image is suffered to fall on a surface precisely as near to the eye as that on which the original colour was displayed, it appears larger or smaller than the original object (22), and this in a rapidly increasing proportion. Lastly, the shape itself soon becomes changed (26).