﻿its Polarization and Colour. 113 



The intensity of the light scattered from a cloud is thus 

 equal to 



A2 (D'-D) 2 • , ST* 

 A D* sm °W 



where ST 2 is the sum of all the squares of T. If T 2 be under- 

 stood to denote the mean square of T (not the square of the mean 

 value of T), and m be the number of particles, 



2T 2 =m.T 2 . 



If the primary light be unpolarized, the intensity in a direction 

 making an angle /3 with its course becomes 



(D'-D) 2 2/P ,mT 2 



Backwards from the cloud the light is thus twice as bright as 

 normally. To the light scattered nearly in the direction of the 

 primary ray our expression does not apply. 



Fig. 1. 



500 H. G F I E D (1 B 1000 A 



Fig. 1 shows the curve representing the intensity of the scat- 

 tered light for each part of the spectrum, referred to the intensity 

 in the primary light as a standard. The abscissa being propor- 

 tional to X, the base line represents the diffraction-spectrum with 

 the principal fixed lines. Over the brighter portion of the spec- 

 trum from B to G the curve differs but little from a straight line, 

 while the small curvature is turned 'downwards, indicating a defi- 

 ciency in the green and yellow. 



Before making out the theory, I had endeavoured to ascertain 

 by observation the actual prismatic composition of the blue of 

 the sky, and had obtained preliminary results. The experimental 

 method (the description of which I must reserve for another oppor- 

 tunity) was fully adequate to the comparison of two given lights ; 

 but the difficulty was to find something to compare the blue 

 light with. In the only complete set of observations that I have 

 hitherto been able to make, the blue of the sky (apparently a 

 very good one) taken from the neighbourhood of the zenith 

 was compared with sunlight diffused through white paper, 



Phil. Mag. S. 4. Vol. 41. No. 271. Feb. 1871. I 



