This apparatus is often known as Maxwell's colour-box, and is in fact a spectroscope reversed. With a collimator and prisms we form a spectrum on the focusing-screen of the camera (Fig. 6), by light coming through the slit, and we can obtain light on the distant screen, a patch of any colour, by placing in the spectrum slits as given at Fig. 30. If we were to illuminate the slits so placed with white light, and look through the slit of the collimator, we should see the front surface of the first prism illuminated by the mixture of the colours which would, when the light illuminated the collimator slit, have formed one colour patch on the screen. In Maxwell's apparatus, the slits S₁, S₂, S₃ are illuminated by the light reflected from a white card C, placed in the sunshine, the rays passing through them fall on two prisms P₁, P₂, are reflected back again through these prisms by a concave mirror M₃, are received on another mirror M, and fall at E on to the eye. At A is an aperture in the box, letting through white light on to a mirror M₁, which reflects it through a lens L on to M₂, which again reflects it on to M, and so to the eye at E. Thus at E an image of the prisms, and an image of the aperture are seen, and the white light of the latter can be compared with the mixture of the colours formed by the prism passing through S₁, S₂, and S₃. Suppose we have one slit S₁, the white light will be decomposed by the prisms, and will be seen at E as light of the same colour as would be seen at S₁, if the light were sent from E to S₁, and so with the other slits. Thus when two or three of the slits are uncovered, the light falling on the eye at E will be a mixture of two or three colours. There are two drawbacks to the mode of illumination used, one being that the quality of sunlight varies, and therefore colour equations will not be accurately comparable one with the other; and the second is that the light reflected from the card is not absolutely the same in all directions, and it cannot be perpendicularly placed to each of the rays which strike the prisms, after passing through the different slits. This latter is a small objection, and is not of much account, but the first drawback is a more serious one. Fig. 37.--Maxwell's Curves of Colour Sensations. With this apparatus, then, Maxwell formed his colour equations, but he fixed as the colours which may be called his standard colours, portions of the spectrum which are certainly not pure, and hence he got curves which are not as perfect as those of Kœnig. It will be seen, for instance, that his red and violet curves do not overlap, but touch each other near E. Were this true, the green colour-blind person should see a dark space in the spectrum, since the green sensation is missing in such eyes. As a matter of fact the luminosity of the spectrum is very considerable to such a person at this point.