Experiments in a Highly Rarefied Gas. 399 



generally accepted, as the opinion is held that the hypothesis 

 of: a unit emission is in contradiction with the laws of optics 

 and cannot explain diffraction and interference experiments. 



Attempt to explain from the unit point of view the 

 Interference experiments hitherto known. 



J. J. Thomson has, however, shown* that all the inter- 

 ference experiments known till now, can be explained by the 

 supposition of corpuscular units — at least qualitatively, — if 

 one assumes that the energy in the light unit remains only 

 invariable as long as this unit does not pass through matter. 

 In passing through matter, ib sets up resonance vibrations 

 amono' the systems of electrons. These vibrators, therefore, 

 take up energy from the primary unit ar.L become them- 

 selves centres for the emission of secondary units. All these 

 secondary units are, among each other and with the primary 

 unit, in fixed phase relation due to resonance. This increase 

 in the number of centres for unit emission together with the 

 phase relation will produce in interference experiments much 

 the same effect as light-waves. 



" If light, however, were produced in a gas at a very low 

 pressure we should expect that in the gas itself the energy 

 in the radiation emitted by a luminous atom or molecule 

 would not be nearly so much subdivided as after it has passed 

 through the glass walls of the tube enclosing the gas, and so 

 that the energy in the light inside the vacuum-tube might 

 be not nearly so uniformly distributed through space as when 

 it got out of the tube. It would be interesting to test this 

 point by seeing whether interference-fringes could be pro- 

 duced as easily and to the same extent in a good vacuum as 

 under ordinary conditions, the light being produced in the 

 vacuum either by the electric discharge or preferably by 

 canal-rays." 



If it is, therefore, possible to make a gas emit light at 

 such a low pressure that an emitted light unit of corpuscular 

 cross-section on its way through the gas has practically no 

 chance of hitting a molecule of this gas, no resonance can 

 come into play. And interference experiments under such 

 conditions will lead to definite conclusions with regard to 

 the resonance hypothesis. 



The whole resonance hypothesis rests, of course, on the 

 supposition that a single light unit alone or a random suc- 

 cession, of units is not able to produce something like inter- 

 ference-fringes, when passing over an interference apparatus. 



* J. J. Thomson, Camb. Phil. Soc. Proc. xvi. p. 643 (1912). 

 2 E 2 



