These numbers are remarkable, and we may enforce what they mean in this way. The energy of radiation, and of light also when of ordinary luminosity, varies inversely as the square of the distance from an incandescent body when of small dimensions. But from the above it seems that a white screen receiving the rays from an amyl-acetate lamp in an otherwise perfectly dark place, and having a colour which stimulates the red sensation alone, would be invisible at 58 feet distance, for there would not be enough energy transmitted to stimulate the red perceiving apparatus sufficiently to give the sensation of light. If it were an orange light, such as sodium, of the same luminosity, we should have to move it from the screen 142 feet before the same result was attained. With the green light at E, the distance would be 550 feet, and with the violet the distance would be increased to 1000 feet. The reduction in intensity of white light, which, when of ordinary brightness, is warm, would make it colder, for the red would disappear, and finally the residue of light, just before extinction, would become a cold grey, due to the absence of all colour. The changes in hue that would occur are variable, the variation being due to the loss of colour of the different rays for different amounts of reduction, and then their final extinction. We can place two patches of white light on the screen, and gradually reduce one in intensity, keeping the other of its original value. No one would expect that the two would be dissimilar in hue, as they appear to be when the former is moderately near the extinction value. If we wish to see this perfectly, we should use an extinction box and view it away from the surroundings, which must be more or less slightly illuminated. It has already been stated that the persistency curve for persons who have normal colour vision is closely the same as that recorded for those who are of the monochromatic type. As this is so, we must expect to find that the persistency curve of these last is the same as their luminosity curve. We put this to the test of experiment and found that our reasoning was correct, for the persistency curve could be almost exactly fitted to it. (See table, pages 217 and 222.) The slight difference between them can be credited to the fact that the whole eye may have been brought into use during the extinction observations, the centre of the eye not being exclusively used. The Figure 31 shows both the extinction and the persistency curves, and also the curve of luminosity for the normal eye. [Illustration: FIG. 31.]