1<S6 BI'.LI. SYSTEM TECHNICAL JOURNAL 



measuring the angle 4> and using equation (46). Let us now try the 

 corpuscle-theory on the problem. 



Putting as heretofore the value h/X for the momentum of the cor- 

 puscles, translate (46) into the language of the alternative theory; one 



gets: 



sin (/) = h/ap. (47) 



In words: a corpuscle of momentum p, passing through any slit, is 

 particularly likely to bend around through an angle (/> of which the sine 

 depends in a certain way on its momentum and on the distance to the 

 next slit. 



Which is to say: the likelihood that a corpuscle entering a slit will 

 bend its course through a certain angle depends on the presence of 

 other slits in the same wall, and on the distance between these slits. 



But the reader will inquire: how does the corpuscle entering one of 

 the slits know that the other slits are there? If all the other slits were 

 suddenly stopped up, the first-order maximum would vanish, the 

 likelihood that the corpuscle would turn in the direction given by (47) 

 would fall to zero; but how could it know that they had been stopped 

 up? 



Well! this is precisely the strange and extravagant property with 

 which we are forced to endow the corpuscles, if we want to use the 

 particle-theory to explain diffraction. It must be supposed that when 

 passing through a slit, a particle of light knows whether there are other 

 slits and, if so, how they are spaced. It must be supposed that an 

 X-ray particle striking an atom in a crystal knows that there are other 

 atoms in a regular array, and knows moreover just the pattern and the 

 scale of that array. It must be supposed that electrons enjoy a like 

 omniscience. Or to express it in more technical language; the prob- 

 ability that a corpuscle of light, of electricity or of matter shall be 

 deflected through a given angle when it strikes an atom or passes 

 through a slit must be supposed to depend on the arrangement of the 

 other atoms or the other slits in the vicinity. This idea is not easy to 

 accept; but it must be accepted, if one is to build up a complete cor- 

 puscular theory of any of these entities. 



But if one accepts it, one finds that the stipulation (47) turns out to 

 be another example of the general quantum-condition of which, in (44), 

 we have already met one instance. For write it thus: 



ap sin (j) = apt = nh, « = 1, 2, 3 . . ., (48) 



the factor n being now introduced to take account of the maxima of 

 second, third, and higher order which also occur on the screen, though 



