CONTEMPORARY ADVANCES IN PHYSICS 73 



quency coincides with a frequency in the spectrum of the scattering 

 substance, it does not necessarily follow that we are dealing with a 

 case of fluorescence. Scattering without change of frequency is very 

 much intensified, when such coincidence is brought about. Ex- 

 perience teaches this, and the wave-theory also; for a vibrator scatters 

 waves most powerfully, when they and it are in resonance together. 

 Scattering with change of frequency may follow the same rule. The 

 proof of fluorescence, then, turns finally on this: can it be shown that 

 there is an interval of time between the moment when the primary 

 quantum impinges on the molecule, and the moment when the second- 

 ary quantum leaves it? 



Often with solid substances one can actually see that the secondary 

 rays continue to emerge for an appreciable time after the primary 

 rays are discontinued; but with gases no such great delay has so far 

 been observed.^ However, there are sometimes indications that be- 

 tween the arrival of the primary quantum and the departure of the 

 secondary quantum, there is an interval of time during which some- 

 thing can happen to the molecule or atom — something which changes 

 the nature of the departing quantum, and may even prevent it from 

 ever being born. Pure rarefied gaseous mercury and sodium and 

 iodine, to take three instances, emit light vividly when they are 

 illuminated; but if they are made very dense, the intensity of the 

 emitted light is much reduced, or its spectrum is entirely changed, or 

 both of these things happen ; so also when they are mixed with gases 

 such as hydrogen or argon. 



Such results, it is clear, are difficult or impossible to explain if the 

 emitted light consists of primary quanta which have rebounded 

 instantaneously from collisions with (say) mercury atoms; for such 

 collisions would be more numerous when the gas became denser, and 

 not much less numerous when the gas was diluted with argon ; and the 

 rebounding quanta would vary proportionately in number, while the 

 frequency-shifts which they display would not be changed (unless one 

 were to hit two or more atoms in succession). However, if the mercury 

 atoms absorb the primary quanta, and hold on to their energy for a 

 while, and subsequently by some independent process release it, all 

 these efTects are quite easy to interpret. The atom which has ac- 

 cepted the energy of a quantum, and has not yet decided to disgorge it 

 in the form of a secondary quantum, may meet another atom and un- 

 load the energy in part or altogether. It is certain that this can 

 happen ; for when mercury vapor is mixed for instance with thallium 



^ Methods whereby it might be possible to measure the time-interval have been 

 suggested by Ruark. 



