But suppose that instead of a red wall or a blue wall we focus our camera upon one which is half red and half blue. Then it is easy to perceive that we shall get a plate which is half one colour and half the other. Moreover, it follows that a wall covered with a mosaic of red, blue and green would give us a plate duly coloured in the same way. But when we go a step further and photograph, say, a landscape, which may contain a vast range of colours, we find a difficulty in believing that they can all be rendered by the simple process of covering or leaving uncovered grains either blue, red or green. It can be done, however, since the other colours may be made up of two or more of these three in varying proportions. For example, should there be something in the landscape of a darker, more blue, shade of green than the green grains, then the light proceeding from that object, while passing freely through the green grains upon which it falls, will slightly penetrate the neighbouring blue ones as well, and so at that point on the plate there will be not only green grains visible, but some of the blue grains partly visible also. The light from the blue grains will enter the eye along with that from the green grains, and by so doing will add just that amount of blue to the green as to give it the right shade. After this manner is the whole picture built up. It is, of course, really a mosaic, consisting entirely of little coloured patches, but since they are so small none can be seen individually, all merging together in the eye so as to form a picture in which colours change imperceptibly from one into another. To sum up, then, what happens is this. We start with a layer of coloured grains; the action of taking and developing the photograph covers up some of these grains and leaves others exposed, and the action of the light is such that those which are left visible produce a picture closely resembling the original, not only in form but in colour. But there is one other interesting point about this process which deserves mention. The differently coloured lights are not of the same power photographically. Red light, as we know well, is very weak in this respect, wherefore, we use it in the dark-room. A faint red light will have no perceptible effect upon a plate unless it be exposed to it for some time. Blue light, on the other hand, is very active, and were the blue and red lights to be allowed to act equally on the autochrome plate, the result would be much too blue. It is therefore necessary to handicap the blue light, as it were, by placing a "reddish-yellowish" screen either just in front of, or just behind, the lens to cut off a proportion of the blue rays. The other very successful process is known as the Dufay dioptichrome process. It differs very little from the Lumière except in detail, the selective screen being formed of small coloured squares instead of by a mass of little grains.