310 FRESNEL OX DOUBLE REFRACTIOX. 



described by the luminous molecules on the emission hypothesis, 

 has the same direction as the ray drawn from the centre of the wave 

 to the point of its surface under consideration. That which w^e 

 have previously said to establish this principle will have perhaps 

 appeared sufficient ; we think it useful nevertheless to support 

 it yet further by a new consideration drawn from another mode 

 of judging by experiment of the direction of the refracted ray. 



New consideration, which shows further that the radius vector 

 of the surface of the wave is really the direction of the luminous 

 ray. 



Suppose, as just now, that the incident wave is plane and 

 parallel to the surface of entry of the crystal, but that the screen, 

 pierced by a small hole, is placed on the fii'st f\ice, instead of on 

 the second ; and that we wish to judge of the direction of the 

 ray refracted through the point D (fig. 12), where the light thus 

 introduced strikes against the second face. The point which 

 will be regarded as answering to the axis of the luminous beam 

 will be the centre D of the small bright and dark rings projected 

 on the face F D ; and it is in this central point that the maximum 

 of light will be found if the hole m« is sufficiently small relative 

 to the distance E D. The position of the centre D is determined 

 ■p- TO by the condition that the rays start- 



ing from the different points m and 

 w of the circumference of the opening 

 arrive at the same time at D. This 

 point must be the most strongly 

 illuminated spot so long as the dia- 

 meter of the opening is sufficiently 

 small with regard to the distance E D for the difference of route 

 between the rays starting from the centre and circumference, not 

 to exceed a semi-undulation. Now, in order to compare the 

 route of the elementary disturbances which emanate from the 

 various parts of the surface of the wave comprised within the 

 extent of the small opening, we must consider the waves which 

 they would produce separately in the same interval of time, and 

 thence conclude the difference between their moments of arrival 

 at D. Let r D 5 be the elementary wave, having for centre the 

 middle E of the opening ; if a tangent plane F D be drawn to it 

 parallel to the incident wave A B, the point of contact D will 

 satisfy the condition just announced ; for the elementary wave 



