748 Transactions of the Society. 



dence upon the mirror. It must be borne in mind, however, that 

 the direction of the reflected rays is influenced by mass and shape 

 of the cloud as a whole, and that its constituent vapour particles do 

 not present a continuous reflecting surface. The numerous minute 

 fields of light and shade which may be observed within a compara- 

 tively circumscribed portion of cloud surface abundantly prove the 

 actual inequality of reflection. 



From the diagrams it may be gathered that one ray only out 

 of each parallel beam occupying the mirror surface with a given 

 degree of incidence actually falls on the object. And, further, that 

 the rays which collectively form the illuminating pencil are singled 

 out, so to speak, by their fulfilment of the necessary condition of 

 converging incidence from a large area of light source. A whole 

 cloud or pile of clouds may in this way be utilized, though the 

 general surface is so unequally bright that the darker portion will 

 frequently reduce the effect of the brighter to below the average 

 intensity of sky light around the cloud. Hence the concave mirror 

 is preferred to the plane, because, acting on a different princi2:)le, 

 it collects the relatively small but bright surface of sunlit cloud 

 without diluting its intensity by including the larger darker 

 portion. But under ordinary circumstances of sky light illumi- 

 nation, the convergence of light upon the plane mirror is not only 

 a necessary consequence of optical law, but also the necessary condi- 

 tion of an adequate illumination. And it follows that the size of 

 the mirror and its nearness to the stage are important points in the 

 design of a Microscope, and equally requiring attention in the 

 practice as in the theory of illumination. 



It has been already noted that a cloud surface is not continuous, 

 like a mirror surface, and that its shape greatly influences the 

 direction of reflected rays. Cloud light is, in fact, self-luminous 

 in the same sense that the light of the firmament is ; that is to say, 

 the solar rays, falling on the cloud, are refracted and reflected and 

 dispersed amongst its own vapour particles. Consequently the 

 parallel incidence of solar rays by no means conditions parallelism 

 or uniformity of the reflected rays. This character of self-lumi- 

 nousness may be contrasted with reflection pure and simple, as, for 

 example, the dazzling glare of a window-pane upon which the sun 

 shines. Nobody would interpret the reflection from a smooth glass 

 surface otherwise than as a simple transference of the sun's light 

 in a new direction, nothing else being changed. The radiation of 

 sun rays is simply continued from the surface (not self-luminous 

 substance) of the window-pane. The cloud-reflected hght is, on 

 the other hand, a residual eflfect, after absorption, refraction, and 

 dispersion of the original light rays have been carried on in the 

 intervening cloud matter to such an extent as to lower the specific 

 intensity of the reflected light beyond calculation. It caimot there- 



