Experiments in a Highly Rarefied Gas. 401 



absorption line of the Hg-vapour at X= 2536. According 

 to Wood (I. c.) its intensity is reduced to the half value by 

 passing through a 5 mm. thick layer of Hg-vapour of 

 0"001 mm. pressure. It shows no trace of polarization. 



There is hardly much doubt that the exciting light in- 

 creases the amplitude of a prominent frequency in the Hg- 

 atom till it emits light, without, however, ionizing the vapour. 

 About the mechanism of this process nothing is known at 

 all. We do not even know what the ratio is of the number 

 of radiating molecules to the total number of molecules for 

 ;a given intensity of the exciting source. 



Considering the limit of pressure, mentioned above, it 

 rseems at first sight that even Hg-vapour at 0*001 mm. 

 pressure is still about ten times too dense in order that its 

 Tesonance radiation may be used for the proposed experi- 

 ment. Fortunately that is not the case. For Hg-vapour, 

 >being a vapour, does not necessarily fill a whole vacuum 

 with uniform density. An experiment, described later, will 

 ■show that saturated Hg-vapour can fill uniformly one 

 chamber, while in a second chamber, in connexion with the 

 first one by means of a narrow slit, the Hg-vapour pressure 

 •can be kept practically at zero, if in this chamber the Hg- 

 vapour is condensed much more rapidly than it is supplied 

 by diffusion through the slit. For our experiment the first 

 chamber, filled with Hg-vapour in resonance radiation, is to 

 be used as source of light ; the second chamber has to con- 

 tain the interference arrangement. The condition that this 

 second chamber must be practically free from Hg-vapour 

 in our case has to be fulfilled for another reason still : the 

 resonance radiation coming from the first chamber would 

 be absorbed entirely before reaching the interference 

 : arrangement. 



If it is possible to excite Hg-vapour of 0*001 mm. pressure 

 in a layer of only 1 or 2 cm. thickness to a sufficiently 

 brilliant light emission, the resonance radiation of Hg-vapour 

 seems to be a source of light for our experiments. 



But before we arrive at this conclusion we have to discuss . 

 the point, already alluded to : As the Hg-vapour is radiating 

 only under the influence of an exciting source of light, 

 which we assume supplies the energy in waves, the question 

 •arises whether these waves perhaps enforce upon the radiating 

 systems of the Hg-vapour some sort of fixed phase relation 

 between the emitted units. 



But this could happen only 



(1) if we would have to assume that the trains of oscilla- 

 tion only come from radiating Hg-molecules in optical 



