324 BL/J. SVST/IM TECIISICAL JOLRNAL 



that radiation traxi'ls around the world in corpuscles of energy hv 

 and momentum hv c, which never expand, or at all events always 

 remain small enough to be swallowed up in one gulp by an atom, or 

 to strike an electron with one single concentrated l)low. 



But it is unfair to close the case without |)leading once more the 

 cause of the undulatory theory — the more so because, in the usual 

 fashion, I have understated the old and presumptively familiar 

 arguments in its faNor, and gi\eii all tlic advantages to the arguments 

 of the opposition, which still Ikim' liic force and charm of novelty. 

 Furthermore, I ma\- ha\e i^roduced the impression that the conception 

 of the qiiantimi actualh' unites the corpuscular theory with the wave- 

 theory, mitigating discord instead of creating it. Why are we not 

 really \-oicing a perfectly competent wave-theor\- of light, when we 

 imagine wave-trains limited both in length and in breadth, so narrow 

 thai they can di\e into an atom, but so long that they contain hv of 

 energy altogether?.A7a;«CH/an' wave-trains, so to speak, like the tracing 

 of a sine-wave in chalk upon a blackboard, or the familiar picture of 

 a sea-serpent.-' 



Well, the dilticuity is ihat I hi' pjicnomcna ol iiUcrtcrence and ot 

 diffraction, which are the basis of the wave-theory, imply that the 

 wave-trains are broad, that the>' ha\e a consitlerable cross-sectional 

 area; these phenomena should not occui-. if \\\v wave-trains were 

 filaments no thicker than an atom, or e\cn so wide that their cross- 

 sectional area amounted to X-'. Let me cite one or two of these 

 phenomena, in tardy justice to the imdulatory theory, as a sort of a 

 makeweight to all the "(|ii.intum " phciminena I have described. 

 Imagine an opatjue screen witji a slit in it; light flows against the 

 screen from behind, some passes through the slit. The slit may be 

 supposed to be half a millimetre wide, or e\en wider. If light consists 

 of (|uanta only as thick as an atonj. or i-\en as thick as the wave- 

 length of the light, they will shoot ihn)ui;ii the slit like raindrops or 

 sand-grains through a wide open sk\lighl. If i1h\' are all mo\ing in 

 jjarallel directions before the>' reach the sHl, tiie\- will continue so to 

 mov-e after the>' pass through it— for how shall the\- know that the slit 

 has any boundaries, since they are so small and the slit is so large/ 

 The beam of light which has passed through the slit will always 

 retain the same cross-section as the slit. But we know that in truth 

 the beam widens after it goes through the slit, and it develops a 

 peculiar distribution of intensity which is accurately the same as we 

 should expect, if the wavefront is ivider than the slit — so much wider, 

 that the slit cuts a piece out of it, which piece spreads outwards inde- 



