ON LIGHT. 329 



mena referable to the head of diffraction, turns on two 

 considerations, one of which may be regarded as a 

 theoretical postulate (founded, however, on the analogy 

 of sound), viz., that if a portion of a luminous wave be 

 intercepted, the non-intercepted undulation spreads 

 laterally into the dark space beyond, diminishing, how- 

 ever, in intensity as the lateral deviation of the ray (or 

 perpendicular to the wave) from its original rectilinear 

 course increases : the other, a natural consequence of 

 the mode in which a wave is propagated, viz., that every 

 point in the surface of the non-intercepted portion maybe 

 regarded as the origin of a new wave spreading out spheri- 

 cally in all directions from that point as a centre ; only with 

 this proviso, that all such secondary waves start, each from 

 its own origin, at the same precise instant of time and 

 in the same precise ^hasc of its luididation. And this, be- 



Fig. 10. 



cause all belong to one wave surface, and are therefore 

 necessarily coincident in time and identical in phase. 



(113.) To show how this last consideration affects the 

 question of diffraction, let us suppose a point p on a 



