450 SCIENCE PROGRESS 



fact that, in addition to the part played by minute dust particles 

 in this connection, the actual molecules of air act in a similar 

 manner, and cause a selective refraction of the light. In this 

 way it was considered that even in the absence of larger 

 particles of matter, the observed properties of sky light could 

 be accounted for. 



Light which has been reflected or scattered by these minute 

 dust particles in the atmosphere or by molecules shows the 

 following properties. The rays which are reflected at an angle 

 of 90 to the incident light are completely polarised, and the 

 blue light is reflected much more completely than the red. 

 It is calculated from theoretical considerations that with different 

 wave-lengths the ratio of the reflected to the transmitted light 

 varies inversely as the fourth power of the wave-length. 

 According to this, blue light should be reflected to about 

 eight times the extent of the red. 



By measuring the intensity of the light which proceeds 

 directly from the sun, Rayleigh's theory enables a calculation 

 to be made of the composition of light proceeding from the 

 sky, if controlled entirely by this principle of scattering. Experi- 

 ments made with an artificially produced turbid medium, such 

 as a steam jet or a fine precipitate suspended in water, show 

 that this relation is very closely followed. Similar measure- 

 ments on sky light can readily be made by means of a spectro- 

 photometer. This instrument is first sighted directly on to 

 sunlight, and a measurement made of the intensities of light 

 in the different parts of the spectrum. These values are then 

 compared with measurements made on light proceeding from 

 the sky when viewed in a direction at right angles to the 

 direction of the sun. According, then, to this theoretical 

 relation deduced by Rayleigh, the intensity is highest in the 

 violet, and falls rapidly towards the red in proportion to the 

 inverse fourth power of the wave-length. 



For wave-lengths still shorter than those in the violet the 

 intensity would be expected to fall again rapidly, on account 

 of the known absorption of the ultra-violet light by the atmo- 

 sphere. This absorption, however, in presence of the admitted 

 constituents of a clear atmosphere should have no appreciable 

 influence in the region of the visible spectrum, so that the 

 relation between intensity and wave-length should be capable 

 of being plotted by means of an even curve. 



