84 



BELL SYSTEM TECHNICAL JOURNAL 



quency shifts observed in and near the visible spectrum. Here is an 

 instance: Landsberg and Mandelstam working with quartz observed 

 frequency-shifts corresponding to infra-red lines of wave-lengths 9ix, 

 13.5;u, 21.5/z, 48/i, and 81;u respectively; there are Reststrahlen of wave- 

 lengths 8.7, 12.8 and 20.7, while the other two wave-lengths cited lie in 

 gaps of the infra-red spectrum unexplored as yet. 



A photograph by F. G. Brickwedde ^"' which I reproduce as Figure 8 

 illustrates this effect. The very broad black band is due to primary 

 light of wave-length about 2536, scattered without change of fre- 



I 55°C 



300° 



525° 



nwiiiiT 



Fig. 8 — Light scattered by a quartz crystal; unshifted line is 2536, lines denoted 

 by S and A are shifted towards lower and higher frequencies respectively. (F. G. 

 Brickwedde.) 



quency; its excessive width is due to the great intensity of this light. 

 The lines marked S21 and S81 consist of quanta which have spent 

 part of their energy in exciting vibrations, those corresponding to 

 the Reststrahlen of wave-lengths 21^ and 81^ respectively. The 

 lines marked A21 and A81 consist of quanta which have received 

 energy from these vibrations. These ' ' Anti-Stokesian ' ' lines evidently 

 grow more intense as the temperature of the crystal is raised, as they 

 should, for the reason which I have already stated: as the substance 

 grows warmer, the percentage of the molecules spontaneously in 

 vibration is increased. In addition, the shifted lines all move inward 

 toward the position of the unshifted light as the temperature rises — 

 one may see this by comparing those on the low-frequency side with 



^^ I am much obliged to Dr. Brickwedde for furnishing me with a print of this 

 photograph. 



