72 BELL SYSTEM TECHNICAL JOURNAL 



say, the primary quanta have exactly the right energy to transfer a 

 molecule from its original state to some other (excited) state. Suppose 

 however that this condition is fulfilled ; then the quantum of fluores- 

 cence-light is emitted when the molecule passes from state B to state 

 A, and therefore has the frequency iiba- But because of the energy- 

 relations, we have 



nBA = flBN — nAN = no — Uan, 



which means that the quantum of fluorescence-light has exactly the 

 same energy as a quantum of the primary light would retain, if it had 

 been scattered from the molecule after communicating to this latter 

 the energy requisite to transfer it from its original state N to the state 

 A. The fluorescence-light is shifted in frequency from the primary 

 line by exactly the same interval as the Raman line corresponding to 

 the transfer of the molecule from N to A. 



It follows then that one can never decide by measurement of wave- 

 length whether a line in the scattered light is due to fluorescence or to 

 scattering with change of frequency. '^ This is inevitable; for in either 

 case the molecule involved in the process starts from the same initial 

 and ends in the same final state, and the frequency of the departing 

 quantum depends on nothing but the difference between these two. 

 The only question at issue is, whether the molecule has gone from the 

 initial to the final state directly, or via the temporary state B. 



One way of solving the question seems obvious: to vary the fre- 

 quency of the light with which the molecules are irradiated, and notice 

 whether the shifted line which is under observation — a line is identified 

 by the amount of its shift, not by its actual frequency, so that a given 

 line travels along the spectrum pari passu with the primary light — 

 makes its appearance when and only when the quantum-energy hna of 

 the primary beam coincides with the energy-difference between the 

 initial state of the molecule and any one of its other states. In other 

 words: does the shifted line appear only when the primary quanta can 

 be absorbed by the molecules — when the infalling light coincides with 

 a line of the absorption-spectrum of the substance? or does it appear 

 always? If the latter is the case, it is the Raman effect which we have 

 before us, at least when the wave-length of the infalling light is such 

 that its quanta are not absorbed. 



If however the shifted line is most intense when the primary fre- 



' This niay be too strongly stated; one might observe a fluorescence-Une emitted — 

 to use the foregoing symbols — by reason of the transition from ^ to iV (not that from 

 B to A) which could not be a Raman line correlated with a transfer of the molecule 

 out of the state N. On the other hand it could be a Raman line associated with a 

 transfer of the molecule out of state A, so that one would have to assess the relative 

 likelihoods of the states A and N among the molecules. 



