Kunz — Electromagnetic .Emissioii Theory of Light. 319 



source of liglit in motion varies with the direction of propaga- 

 tion of liglit. In tlie direction in which the source of light is 

 moving, the velocity of light is equal to c + v, in the opposite 

 direction equal to c — v, in the perpendicular direction, c. 



Tlie experiment should be able to decide between these 

 alternative theories. Am'ong different possibilities the canal 

 rays lend themselves perliaps in the first place to an experi- 

 mentum crucis. We are here dealing with positive particles 

 whose velocity in hydrogen may be as high as lO'cm./sec, or 

 only 300 times smaller than the critical velocity, c, of light. If 

 it could be shown that the velocity of lighr, given out by 

 those moving particles is the sum c + v, the principle of 

 relativity would disappear and the electromagnetic emission 

 theory would be victorious. If the velocity of these positive 

 particles was found to be c in all directions, the emission 

 theory would have to disappear forever. 



The velocity of the source of light in the canal rays deter- 

 mined by means of the Doppler effect varies from 30-80 per 

 cent of the velocity computed from the cathode fall of poten- 

 tial and the value e/m. A simultaneous determination of the 

 velocity of canal rays by means of the Doppler effect and by 

 means of the electrostatic and electromagnetic deflection would 

 be very desirable. 



§ 4. Interference. 



The emission theory suggests the possibility of a difference 

 in the interference of two beams issuing under a large angle 

 from the same source and the interference due to two beams 

 issuing in the same or very nearly the same direction. The 

 light from two sources is not susceptible of interference, 

 while interference occurs in a beam from one source, because 

 there are phase relations between the oscillations in the Fara- 

 day tubes, issuing from one source. It may be that for this 

 reason N. Campbell* has not yet succeeded in searching after 

 a difference in the " events " of the photoelectric effect, due to 

 two beanis separated by interference. 



If, however, the energy travels out from the source along 

 certain directions, we should finally be able to detect, for in- 

 stance in the case of Rontgen rays, a scintillation effect, or at 

 least discontinuities in the phosphorescence of a sensitive 

 screen. It is possible, too, that one screen could show such 

 discontinuities while another does not, because if a molecule is 

 made a source of light, it might affect the neighboring mole- 

 cules to emit light also ; and in this way a discontinuity, 

 originally present, might be masked. Perhaps the idea of 



*N. Campbell, Discontinuities in Light Emission. Proceedings of the 

 Cambridge Philosophical Society, vol. xv, p. 310, 1909. 



