of Light and its Theory. 261 



sented either as attracting or repelling the light, it is difficult 

 to explain why the surface of a refracting medium attracts one 

 portion of the light, while it repells another proportion, and 

 even the auxiliary hypothesis of fits of the particles of light 

 to facilitate refraction (or reflection) leaves yet many difficul- 

 ties, when it is considered that one and the same refracting 

 surface reflects more light in proportion as the angle of inci- 

 dence of the received rays is greater, and that every refracting 

 substance, when the angle of incidence approaches to 90°, re- 

 flects almost all the light falling upon it. 



That a level but otherwise rough surface, which irregular- 

 ly disperses the light which falls vertically upon it, should, at 

 great angles of incidence, reflect the rays regularly, is very 

 simply explained by the interference. The colours of mo- 

 ther of 'pearl are of the same nature as those which are pro- 

 duced by the reflection of light from the surface of a system 

 of lines upon glass. Even if they were not by other means 

 easily recognised as such, Dr Brewster's discovery, that a 

 good impression of mother of pearl shows the same colours, 

 and that, consequently, the cause lies in the surface, would 

 prove it. 



It consists also of particles or layers, which in one direction 

 are larger than 6>;, and their magnitude may be very nearly 

 derived from the angle under which the colours of any sort 

 are seen. 



Even those who do not adopt the system of undulations, 

 will acknowledge, if they consider the results of the experi- 

 ments in themselves, that w is a real absolute magnitude. 

 Whatever meaning may be attached to this magnitude, it must 

 follow in every case of that nature, that one-half of it, in 

 reference to the eff"ect, is opposed to the other half, so that if 

 an anterior half combines accurately with a posterior half, or 

 intersects it in this manner under a small angle, the effect 

 ceases, while it becomes doubled, if, for instance, two anterior, 

 or else two posterior, halves combine in one sense. This, 

 laid down as fundamental to interference^ will remain unshaken, 

 because through it alone these extraordinary varied pheno- 

 mena, which are so capable of accurate definition, can be satis- 

 factorily explained. 



It is very probable that in the sequel experiments will 



