3. When light is reflected by an opaque white body, it is not modified in proportion to the variously coloured rays which constitute white light; but, 1. _If the body is not polished_, every point of its surface is to be considered as dispersing the white light which falls upon it, in all directions, into the surrounding space; so that the point becomes visible to an eye placed in the direction of one of its rays. We may easily conceive that the image of the body, in a given position, is composed of the sum of the physical points, which send to the eye so placed, a portion of the light which each point radiates. 2. _If the body is polished_, as, for example, the surface of a mirror, a portion of the light is irregularly reflected, as in the preceding case; while another portion is regularly reflected, giving to the mirror the property of presenting to an eye, suitably placed, the image of the body which sends its light to the reflector. One consequence of this distinction is, that if we regard two plane surfaces which reflect white light, and differ from each other only in polish, it will happen that where the unpolished surface is visible, all its parts will be equally, or almost equally, illuminated; while the eye, when in a position to receive only the light which it reflects irregularly, will receive very little light from the polished surface; but it will receive much more light when in a position to receive that which is regularly reflected. 4. If the light which falls on a body is completely absorbed by that body, so that it disappears from sight, as in falling into a perfectly dark cavity, then the body appears to us black; and it becomes visible only because it is contiguous to surfaces which reflect or transmit light. We know of no bodies which are perfectly black, and it is only because they reflect a little white light that we judge they have relief, like other material objects. 5. When light is reflected by an opaque coloured body, there is always a reflection of white light, and a reflection of coloured light; the latter is owing to the fact that the body absorbs or extinguishes within itself some of the coloured rays, and reflects the others. It is evident that the _absorbed_ coloured rays are of a different colour from the _reflected_ coloured rays; and farther, that if these be reunited with the former, white light will be reproduced. It is evident, also, that unpolished opaque bodies reflect irregularly white light, and the coloured light which makes them appear coloured; and that those which are polished reflect irregularly a portion only of these two lights, while they reflect regularly the other portion.