334 ON LIGHT, 



amount of concordant and no discordant rays, and the 

 illumination will be a maximum. But if p approach 

 nearer, the difference in question will be greater than a 

 semi-undulation, and the portion of an aperture near the 

 edges will send rays to P more or less in discordance with 

 those from the centre, and these will destroy a por- 

 tion of P*S illumination, p approaching this will go on 

 till the difference amounts to an entire undulation 

 (i-5o,oooth) j that is to say, till the aperture extends (as 

 respects the new situation of P) over the whole of the two 

 first zones (A) and (B), of which the second (B) destroys 

 almost the whole of the light from (A) (being equal in 

 area, and differing very little in obliquity). Here then 

 the illumination at P will be nil or very small. P still 

 approaching, the third zone (c) (which sends vibrations 

 in unison with (A) ) will begin to be included within the 

 limits of the aperture, and the illumination will again in- 

 crease to another maximum, viz., when the three, (A), 

 (B), (c), are just included ; thence again it will diminish 

 to a degree of obscuration not quite so complete as be- 

 fore; and so on. Thus as the screen approaches the 

 aperture, its central point, after attaining a maximum of 

 illumination, will suffer a succession of eclipses or obscur- 

 ations gradually less and less complete, with intermediate 

 recoveries, just as we have seen above, is really the case. 

 When the light is not homogeneous, the different coloured 

 rays having different wave-lengths, the obscurations of 

 one colour will not correspond to those of the others, 

 and thus will arise a succession of colours at the central 

 point of the screen, agreeing with those there described. 



