﻿and Absorption by Resonating Gas Molecules. 695 



graph was now made as narrow as possible and five exposures 

 were superposed of five seconds each, the direct light from 

 the arc being screened from the slit. The spectrum showed 

 only the 2536 line, exceedingly narrow and sharp, thus 

 proving that we are dealing with a very beautiful case of 

 pure resonance radiation. In fact, I suspect that this radi- 

 ation will be found to be the most homogeneous which we 

 have, for the vapour is not only at a lower pressure than is 

 usual in vacuum-tubes, but it is at room temperature. In 

 PI. XI. fig. 2, the third spectrum is that of the resonating 

 vapour, the fourth spectrum that of the mercury arc. 



I next investigated the effect of raising the temperature 

 of the tube which was mounted in a small air-bath. It was 

 found, as was to be expected, that, as the temperature rose, 

 the cone of emitted light became shorter and brighter, until 

 it finally disappeared, the emitted light coming from the 

 inner surface of the plate where the incident beam of light 

 entered. As no especial significance is attached to the 

 records obtained in this way, I shall postpone any further 

 discussion of the effects of increasing the density of the 

 vapour until 1 take up the subject of the transition from 

 diffuse scattering of the emitted light to its recombination 

 into a regularly reflected wave, which can be accomplished 

 by increasing sufficiently the density of the vapour. 



We will now take up the important subject of the amount 

 of energy diverted from the incident beam by the resonating 

 gas molecules. 



Amount of Energy diverted from the Primary Beam. 



It is clear at the outset that if we wish to determine the 

 amount of energy diverted by the resonators when they are 

 in exact synchronism with the light- waves, it is useless to 

 make observations upon the intensity of the light after it has 

 suffered transmission through the vapour, even if we are 

 dealing with what we are accustomed to call monochromatic 

 light. All spectrum lines have a finite width, and the par- 

 ticular frequency scattered by the resonating molecules may 

 constitute but a small fraction of the total energy of the 

 spectrum line used to excite the vapour; in other words, it is 

 only the centre of the line that is effective in exciting 

 resonance, the edges of the line not being reduced in 

 intensity by the transmission through the gas. What we 

 wish to determine is the reduction in intensity of that 

 portion of the line, or in other words t\\e, frequency, which 

 is capable of exciting the natural period of vibration of the 



