428 Mr. J. C. McConnel on the 



plane of diffraction, i. e. the plane through the sun and the 

 eye. The breadth of the representative plate or slit, measured 

 on this plane, determines the diffraction of the various spectra, 

 and is the greatest breadth of the section of the filament 

 measured across the sun's rays. 



In fig. 1 is given the case of a hexagonal filament, whose 



Fig 1: 



B 



minor diameter is perpendicular to the " reflecting plane," 

 when the angle of diffraction is 10° ; first when the axis is at 

 right angles to the plane of diffraction, and secondly when it 

 is inclined at only 20° to that plane. 



AS, BS are drawn towards the sun, AE towards the 

 observer, and AB is the breadth of the representative slit. It 

 will be noticed how slightly AB is altered even by this ex- 

 treme inclination. If the major diameter had been perpen- 

 dicular to the reflecting plane, there would have been no 

 alteration. The difference between the major and minor 

 diameters themselves is of somewhat more importance, but 

 each differs by only about 7 per cent, from the mean. 



Among the numerous crystalline forms met with in falling- 

 snow, and therefore presumably in the clouds, there is none 

 approaching to a sphere. But thin hexagonal filaments are 

 common, either separate, or forming part of more elaborate 

 structures ; so it is these which we must regard as the 

 " active principle " in the coloured clouds. Let us then 

 consider the spectacle to be expected in a sky covered with 

 thin clouds composed of filaments of one uniform diameter a. 

 The following table, deduced from figures quoted by Verdet*, 



* Lemons cCOptique Physique, tome i. §§ 69, 77. 



