white light is deprived of at least one constituent sensation which is perceived by normal eyes. This, in all probability, renders the white less luminous to them than those possessing normal vision, so that the comparisons of luminosities are referred to different standards. It may seem a very simple matter to ascertain the correct scale, but it is not, except by the extinction method, which will be described later. At one time General Festing and myself tried to obtain a comparison by finding the limiting illumination at which a book could be read. We got results, but for the purpose in question the values are not conclusive. What we really were measuring was the _acuteness of vision_ in different coloured lights. As a good deal depends upon the optical perfection of the eyes under examination, besides the illumination, we must be on our guard, even if there were nothing else against the method, against taking any such measures as being conclusive. CHAPTER VIII. Before quitting the measurement of luminosity, it may be as well to see whether the curves described are the same whatever the brilliancy of the spectrum may be. We can easily experiment with a very reduced brightness. Upon the screen we have an ordinarily bright spectrum. As the slit, through which the white light forming the spectrum comes, is narrowed, there is an evident change in the relative brightness of the different parts, though the energy of every ray must be proportionally reduced. The red is much more enfeebled than the green, and in brightness the green part of the spectrum looks much more intense than the yellow, which is ordinarily the brightest part. This we have assured ourselves of not only by casual observation, but also by direct measurement. Perhaps I can make this even more decisive to you. Two slits are now in the ordinarily bright spectrum, one in the red, and the other in a green which is near the E line of the solar spectrum. Instead of using one lens to form a single colour patch of mixed light, two parts of a lens, appropriately cut and of the same focal length as the large combining lens, are placed in front of the slits, one bit of lens before each. This artifice enables us to throw the patch of red on one white surface, and the patch of green light on another adjacent to it. By opening or closing one or other of the slits the brightness of the two patches of light are so arranged that there is no manner of doubt but that the red is the brighter of the two. The absolute energies of the rays forming each of the patches are proportionally reduced by closing the slit of the collimator, as before. At one stage both patches appear of about the same intensity. This might be taken for an error in judgment, but to make the change that takes place perfectly plain to you, the rotating sectors are introduced in front of the two slits, and the rays now pass through them. The apertures of the