ON LIGHT. _ 365 



posterior surface which wWXJirsf be sti'uck by the refracted 

 wave. The portions beyond on either side, i n and o m, 

 will sithsequently receive the divergent undulations, which 

 (as we have already explained) give rise to diffracted 

 frinojes bordering the shadows of the screens e f, g h. 

 Thus we see that the space between n and o, and not 

 that between k and l, will receive the full illumination 

 from the aperture f g, which has therefore been propa- 

 gated obliquely in the direction of the lines F n, g o, and 

 not of the perpendiculars f k, G l.* 



(144.) The deviation of the refracted ray from the plane 

 of incidence, and from that of ordinary refraction, will 

 be readily understood when it is borne in mind that 

 whether at a perpendicular or an oblique incidence, a 

 plane exterior wave is transformed by the extraordinary, 

 as well as the ordinary refraction, into a plane i/iterior 

 one, and that the plane of incidence of a 7-ay is perpen- 

 dicular to both these planes. It cannot therefore con- 

 tain the extraordinary refracted ray (which is a radius of 

 the spheroid) without containing at the same time a 

 normal to the elliptic surface of propagation at its point 

 of contact with the interior plane wave, that is to say, 

 unless it contain also the axis of the spheroid. In other 

 words, the extraordinary ray will always deviate from the 

 plane of incidence, unless in the case when that plane 

 coincides with some one of the meridians of the spheroid 

 in question. 



(145.) Since there is no double refraction in the direc- 

 tion of the axis of the rhomboid, it follows that in that 

 * This is Huyghens's explanation, and the correct one 



