[Illustration: _Fig. 34.--Illusion of Luminosity._] The fact is well illustrated by a familiar experiment. Fig. 34 is photographed from a transparency made by superposing several different lengths of gelatine film so as to form a series of steps. At the right-hand end of the image the light has passed through only one layer of the film; in the next division it has traversed two layers, in the next, three, and in the last, four. The luminosity of each of the four squares into which the oblong is divided is, in a physical sense, quite uniform, but the mental standard of luminosity varies for different parts of the image, increasing or decreasing, as the case may be, not _per saltum_, but smoothly and continuously, with the result that each square looks brighter towards the left than towards the right. The appearance, which is often likened to that presented by a fragment of a fluted column, is equally well shown when the diagram is illuminated instantaneously by an electric spark, and cannot, therefore, be accounted for by retinal fatigue. If the squares are separated from one another by distinct lines of demarcation, however fine, the standard of luminosity becomes uniform for each square, and the illusion vanishes. This fact sufficiently disposes of the hypothesis which has been advanced to the effect that the phenomenon is due to physiological causes. I now propose to discuss a curious consequence of the fluctuation of unaided judgment as regards the purity of a colour. When any colour occupies a predominant place in the field of vision, we are apt to consider it as being less pure, or paler, than we should if it were less conspicuous, our standard of whiteness tending to approximate itself to the colour in question. For the sake of clearness let us first confine our attention to a definite colour--say red. An absolutely pure red is one that is entirely free from any admixture of white; in proportion as it contains more and more white, the more impure, or in other words, the more pale does it become, until at last all trace of perceptible redness is lost and the colour is indistinguishable from white. [Illustration: _Fig. 35.--Illusion of Colour._] A convenient way of picturing the scale of purity is shown in Fig 35. The shaded oblong may be supposed to represent a painted strip of cardboard or paper. At the extreme right hand end the colour is supposed to be absolutely pure red; towards the left the red gradually becomes paler or more dilute, and at the middle of the diagram it has merged into perfect whiteness. The figures 0 to 100 from left to right denote the percentage of free red contained in the mixture at different parts of the scale; the luminosity is supposed to be uniform throughout.