“Let AB (fig. 95) be a white opaque surface, _a_, a luminous body, and _m_, a black opaque body, then the shadow _b′_ cast on AB, will appear deeper when observed from P, than as seen from Q. This is a fact which may be easily verified, and the cause of which is easily conceived. In fact, the surface AB, although it disperses the light, must still reflect more of it, in the directions in which the regular reflection takes place; and hence the rays which are reflected round about the shadow, must have a greater intensity in the direction of P than in that of Q, and, consequently, the shadow _b′_ must appear deeper from the point P than from Q. [Illustration: Fig. 96.] “If we now place (fig. 96) two lights in front of the screen AB, at such distances that the two shadows _a′_ and _b′_ should have equal intensities, it is evident that if the eye be placed at P, the shadow _b′_ must appear more intense than the shadow _a′_, and that the reverse will take place if the eye be at Q. But the difference which is then observed, arises not only from the difference in the brightness of the parts surrounding the shadows, but also from a difference in the intensity of the shadows themselves; for the shadow _b′_ is illuminated by _b_, and radiates much more towards Q than towards P; and, on the contrary, the shadow _a′_, which is illuminated by _a_, radiates much more towards P than towards Q. We perceive also why the _differences_ of the tints increase with the separation of the two shadows, and why they become very small when the shadows touch each other; it is because, in proportion as the shadows are farther apart, each of them is illuminated more obliquely, and a greater quantity of light is radiated (by reflection) in the regular direction. When they touch each other, on the contrary, they are illuminated almost perpendicularly, and consequently the shadows radiate light almost equally on either side. “Those anomalies of a like kind which are observed when the shadows are viewed through a translucent body, such as paper or linen, may be referred to a similar cause. We know, in fact, that, in looking through a translucent medium, we always, more or less, distinctly perceive the luminous body behind it, and, also, that there is a very large proportion of the rays which traverse the body, which stray but a little from the direction which they would follow if the substance were absolutely transparent. Consequently, the space which surrounds the shadow is more luminous in proportion as we come nearer to the direction of the shadow; and as the absolute intensity of the shadows diminishes as we come nearer to the direction of the rays which light them, those two effects concur to increase the intensity of that shadow to which the eye is nearer.