74 BELL SYSTEM TECHNICAL JOURNAL 



vapor and the mixture is bombarded with quanta which mercury atoms 

 can absorb and thallium atoms cannot, we nevertheless presently find 

 the thallium atoms possessed of some of the energy which was sent in. 

 Collisions of atoms occur more frequently, the denser and the warmer 

 the gas or the mixture of gases becomes ; the opportunities for diversion 

 of energy, which otherwise would be re-radiated as fluorescent light, 

 become correspondingly more numerous. When therefore the light 

 emitted by an illuminated gas changes its spectrum or fades away as 

 the gas is densified or contaminated, the probabilities are that it is true 

 fluorescence-light.^ The influence of a magnetic field upon the charac- 

 ter of the emitted light may also furnish evidence. 



One sees therefore that the distinction between scattering and 

 fluorescence is by no means immediate. Even the case which seems 

 most explicit of all^ — where the primary light agrees with one of the 

 spectrum-frequencies of the atom and the secondary light is unshifted, 

 as when rarefied sodium vapor is illuminated by one of the D-lines and 

 re-radiates it — is not exempt from doubt. Very likely part of the re- 

 radiated rays is fluorescence-light and part consists of scattered quanta. 

 It is obvious why Raman thought at first that he was observing 

 fluorescence. Others very likely had already noticed the Raman 

 effect, and classified it merely as another instance of the already well- 

 known phenomenon. 



I will now relate some of the details of the recent experiments which 

 have suggested that light may actually be scattered with change of 

 frequency. 



The Raman Effect 



The scattering of visible ^^ light with change of frequency was first 

 discovered by a man who was working with molecular liquids. It is 

 interesting and instructive to consider why the effect, so obvious under 

 these circumstances, had eluded the numerous and notable physicists 

 who had studied — exhaustively, it was thought, — the influences of 

 light on gases and of gases on light. 



In the first place, a liquid contains many more molecules per unit 

 volume than a gas, and therefore offers many more opportunities for 

 collisions of quanta with molecules. This is essential, for collisions 

 which result in excitation of the molecule and in scattering of the quan- 



^ I suspect that Saha, in concluding that the resonance-spectra of vapors dis- 

 covered by Wood are actually examples of scattering with change of frequency, 

 did not take sufficient account of some of these phenomena. Consider for instance 

 those observed by Wood and Loomis {Jour. Franklin Inst., 205, pp. 489-495). 



'" I will reserve the name " Raman Effect " for the scattering with shift-of-frequency 

 of light of the visible and adjacent ranges of the spectrum, as in the X-ray region the 

 effect was earlier discovered. 



