Fig. 11 is a double pendulum arranged as shown. The pendulum A is heavily weighted, whilst the pendulum B is light, being only a string with a small weight attached. This difference in weight was made designedly, to prevent any great effect of the movement of B being shown on A, though that of A must necessarily exercise a great influence on B. The two pendulums are now of the same length. A is set in motion, and as it swings, B also begins to swing, and soon is oscillating with greater motion than A, and continues to do so. The length of the pendulum B is next shortened, and A is again set in motion. B takes up the motion, and increases its swing more and more, but now the two pendulums are in opposite phases, and the motion of A tends to diminish the swing of B, and continues to do so till, after an interval of time, B is once more at rest, when it again will start swinging. The fact is, that when A commences to swing, B also commences; and as long as B and A are moving in the same direction the impulses tend to make B increase its swing, but when they are moving in the opposite direction, or rather, perhaps it should be said, when A begins to start from the highest point of its swing downwards whilst B is travelling upwards, the swing of B will gradually diminish. This, of course, must happen when B is shorter or longer than A, since their times of oscillation are then different. We can now picture to ourselves that when in the perceiving apparatus in the retina the moving parts--probably molecules or atoms--arrive at a certain amplitude, there is then an impression of light, and that it is quite possible that not only those waves whose motion is exactly of the same period as that of the apparatus will set them in motion, but also those waves which are actually of a very different period. If such be the case, it can be seen that waves of light of some periods may set each of the three kinds of perceiving apparatus in motion, and that possibly the resulting impressions given by the sum of all three for a wave out of tune with any of them may be even greater than when the wave period is absolutely the same as one of them. For in the last case a maximum effect may be produced on one apparatus, and the effects on the other two may be insignificant; whilst in the first case the effects on two of them may be so large that their combined effects may have a larger value. [Illustration: FIG. 12.]