68 BELL SYSTEM TECHNICAL JOURNAL 



ventional atom-models.^ Models however can be devised which 

 account for modulation, and perhaps they will become more popular.* 

 In one simple case related to the Compton effect, forced vibrations 

 result in waves which do not coincide in frequency with the primary 

 waves; this is the case of a free electron, which the magnetic force in the 

 light-stream pushes more and more rapidly forward as the electric force 

 makes it swing more and more rapidly crosswise; and as the electron 

 gains speed, the frequency of the waves which it sends to a stationary 

 observer steadily sinks. 



To the thoroughgoing advocate of the corpuscle-theory, however, the 

 problem of the scattered light of shifted frequency seems simple; 

 or, at all events, the first step in explaining it seems obvious. Fre- 

 quency of light, when multiplied by the universal constant h, is the 

 measure of the energy of the corpuscles of the light. Change of 

 frequency therefore means transfer of energy. If a quantum of 

 frequency Wo flies onto an atom and a quantum of frequency Wi flies 

 away, energy in the amount /j(wo — Wi) stays behind with the atom. 

 If fii is greater than Wo, as sometimes happens, the departing quantum 

 takes with it some energy which belonged to the atom as well as all 

 that was brought by the oncoming quantum. As yet there is no 

 picture of the process by which the energy is passed between the 

 matter and the light. But we are not supposed to ask the quantum- 

 theory for such pictures. Perhaps one reason why it seems so much 

 stronger than the wave-theory is, that of the latter we have expected 

 so much more. 



However, visualizable or not, the corpuscle-theory implies that if 

 the scattered light differs in frequency from the infalling light, indi- 

 vidual molecules or atoms are receiving or giving energy in quantities 

 equal to the frequency-difference multiplied by h. Not any and every 

 amount of energy may be annexed or ceded by a molecule or an atom — 

 only certain sharply definite, distinct and separate amounts, equal to 

 the energy-differences between the state in which the particle happens 

 initially to be, and one or another of its various other "permitted ' 

 stationary states. There are exceptions to the rule, as I will state 

 immediately; but in experiments performed with visible light they 

 are not apparent. Correspondingly the frequency-shifts of the 

 scattered light are limited to certain distinct and separate values; 



^ Thus an electron which is subject at once to a quasi-elastic restoring-force, a 

 sinusoidal electric field, and a constant magnetic field perpendicular to the electric 

 field, describes a fixed orbit with a single frequency equal to that of the electric force, 

 and there is no modulation. Mistakes in this respect have been made by various 

 people who theorized about the Wood-Ellett effect. 



■• Such models have been devised by Hartley and by Kennard. 



