sensitive pendulums which are capable of responding to the beats of waves of light, it requires no great imagination to see that one such pendulum will respond not only to that wave of light which is isochronous with it, but also with waves shorter and longer than that particular wave. The same pendulum indeed may respond to the whole of the visible spectrum, but when far off from the maximum the response would be very small indeed. We may therefore assume that though each pendulum may have its maximum increase of oscillation at one part of the spectrum, yet at other parts not only it alone answers to the beating of the waves, but that the other pendulums are also affected by the same, and thus the whole spectrum is recognized by the swings more or less long, of either one, two, or of all three. To Thomas Young is usually attributed the three-colour theory, though it seems to have been promulgated in an incomplete state some time before; Clark-Maxwell and Helmholtz revived it in later years, and it is usually known as the Young-Helmholtz theory. It should be remarked that the three fundamental colour sensations are not of necessity the same sensations as are given by the three primary colours, as we shall see further on. The following figure (Fig. 34) is taken from Helmholtz's physiological optics, as diagrammatic of the three sensations. Fig. 34.--Curves of Colour Sensations. To this diagram there is an objection, in one respect, viz. that it gives the same luminosity-value to the blue of the spectrum as it does to the red and green. It has been seen that if we call the luminosity of the yellow 100, that of the blue is about 5. The objection does not hold if it is remembered that the three maxima of impressions are taken as equal. If the ordinates were increased, so that the maxima were of the same height as that of the photographic curve, the resemblance between them and this last would be very marked. It will be noticed that each of the three colour sensations is not only excited by a limited portion of the spectrum, but by all of it, the height of the curves being a measure of their response.