CONTEMPORARY ADVANCES IN PHYSICS 71 



reason to foresee — then the aforesaid bands would be reduced to lines, 

 shifted from the primary line through frequency-intervals equal to the 

 ionizing-potentials divided by h. Such lines were observed in the 

 early spring of 1928 by B. Davis and D. P. Mitchell, in the spectrum of 

 X-rays scattered by graphite. 



There is one more way in which corpuscles of light can dispose of 

 part of their energy — in setting into vibration atoms which are built 

 into the structure of crystal lattices. Many crystals behave as if they 

 contained oscillators having natural frequencies of the order 10^^-10^'*, 

 and able to emit light of the corresponding wavelengths, which are in 

 the infra-red region of the spectrum. Some of the quanta which strike 

 such a crystal lose energy in being scattered, and the energy which 

 they lose is equal to h times one or another of these oscillation-fre- 

 quencies. This effect was discovered independently in the late winter 

 or early spring of 1928 by G. Landsberg and L. Mandelstam, and by 

 C. V. Raman and K. S. Krishnan. 



Presently I will quote in more detail the data which establish all 

 these facts; but first it is urgent to point out that there is another 

 phenomenon in nature, a phenomenon long known and well known, 

 with which the scattering of light with altered frequency can readily be 

 confused; indeed it is often difficult, and I suspect that it may some- 

 times be impossible, to tell whether in an actual case we have the one 

 or the other before us. I refer, of course, to fluorescence. The de- 

 scription of fluorescence, indeed, reads exactly like the description of 

 scattering of light with change of frequency. Light of one frequency 

 falls upon a substance, and light of another frequency emerges from it. 

 How then shall we discriminate between the two? 



According to the ordinary conception of fluorescence — a conception 

 which has attained to the rank of a definition — the molecule or the 

 atom absorbs a quantum of the incident light, and is put thereby into an 

 excited state; and after a longer or a shorter time, it passes spon- 

 taneously into a state different both from the excited and from its 

 original state, and emits a quantum which is not of the same frequency 

 as the one which it absorbed. (I am considering fluorescence of gases 

 or of dilute solutions, where one can suppose that the quanta are ab- 

 sorbed or emitted by individual molecules ; more complex cases are too 

 complex, for the time being.) Let the original state be symbolized 

 by TV, and the final state by A, and the temporary state by B\ 

 denote by Ean = hnAN the energy-difference between A and N, 

 positive if the energy in state A is the greater; use the letters Ubn and 

 7iBA correspondingly, and denote by «o the primary frequency. Then 

 there will be no fluorescence at all unless «o = i^bn ', unless, that is to 



