1S4 BULL SYSTL.M TliCIINICAL JOURNAL 



liriiim, considered as a gas of which the atoms are corpuscles of hght; 

 while Fermi. Dirac and Sommerfeld between them used it to make a 

 powerful theory of the free negative electricity in metals, conceiving 

 this alternatively as a gas of which the atoms are electrons, and a 

 system of stationary waves enclosed within the surface of the metal as 

 in a box with reflecting walls. 



Diffraction' of Waves axd Diffraction' of Corpuscles 

 The effect of a diffraction-grating upon a beam of light projected 

 against it has always been considered the most striking evidence that 

 light is of the nature of waves and not of corpuscles. Indeed it is 

 considered to suffice in itself to prove the corpuscle-theory untenable. 

 With any common understanding of the term corpuscle-theory, 

 this statement is correct; but we had better put it in the softer form, 

 that the effect of a diffraction-grating on a beam of light proves that if 

 we adopt a corpuscular theory we must endow the corpuscles with some 

 very strange property which nobody ever thought that particles could 

 possess, and which may even seem to be in contradiction with their 

 nature. W'e had better put the statement in this milder way, because 

 it now is known that in spite of all the evidence for individual electrons, 

 a beam of negative electricity is affected by a grating in much the same 

 way as a beam of light. 



Take then almost the simplest conceivable case of diffraction; a 

 plane-parallel beam of light falling perpendicularly on a wall containing 

 many equally-spaced parallel slits, and a part of the light passing 

 through the slits to a screen infinitely far away. On this infinitely- 

 distant screen — which may in practice be brought up to a convenient 

 nearness, b}- means of a lens — one sees a peculiar pattern of light and 

 shade. I single out one particular feature of this pattern : the fact that 

 there are maxima of illumination along certain lines parallel to the 

 slits. One of these, for instance, is straight ahead from the slits, along 

 the direction of the incident beam prolonged; another is oft' to one side, 

 in a direction making a certain angle (say </>) with that of the incident 

 beam; another is equally far off to the other side. These two last- 

 named, the Jirst-order maxinia, are those we shall consider; it will be 

 enough to speak of one. 



By the wave-theory, a first-order maximum is explained as follows. 

 Each of the slits is the source of a secondary wave train of spherical 

 wave crests, stimulated by the primary wave train, and having the same 

 frequency and wave-length. Consider any two adjacent slits. 

 Secondary wave crests start from the two at the same moment. At 

 any point equally distant from the slits, they arrive simultaneously, and 



