PHOTONS AND ELECTRONS 33 



From my description of the mechanism of the production of Ught by 

 light, the reader will infer that the produced Hght is always of lesser 

 frequency and greater wave-length, or else of the same frequency and 

 wave-length, as the producing light. This is a nearly but not quite uni- 

 versal rule of experience; it was discovered long before the advent of the 

 current notions of light and atomic structure and bears the name 

 "Stokes's law." 



The exceptions to Stokes's law do not invalidate these current notions 

 but are brought about by interesting special causes. With molecular 

 gases there is no single state sufficiently distinguished from its neighbors 

 to be called "the normal state." Instead, the state of lowest energy may 

 have dozens of neighboring states, whereof the energy values exceed its 

 own by so little that in a gas at room temperature there are great numbers 

 of molecules in each of them. Say that in order of increasing energy 

 these states are denoted by A^o, Ni, N2, .... It is then possible for the 

 following pair of events to occur: a molecule in, say, state A^2 absorbs a 

 photon of energy hv and is thereby transferred into an excited state A, 

 from which it subsequently makes a spontaneous transition into the 

 state No. The photon emitted in this second event has an energy which 

 is greater than that of the original photon by the amount of the energy 

 difference between the states A^o and A^2; the produced light is of higher 

 frequency than the producing light, and there is an exception to Stokes's 

 law, sometimes called an " anti-Stokesian " line. Exceptions of another 

 type will be mentioned presently. 



Another interesting phenomenon occurs when the irradiating light is 

 very intense. Imagine an atomic gas, and consider four of the stationary 

 states of its atoms: the normal state N and excited states A, B, C. Sup- 

 pose the gas illuminated by light of two frequencies simultaneously: the 

 frequency NA — by which I mean that of which the photon energy is just 

 equal to the energy difference between N and A — and the frequency 

 AC. The photons of the former will transfer the atoms which they 

 strike into the state A. The photons of the latter will have no effect on 

 normal atoms; but if one of them happens to strike an atom which is in 

 the state A by virtue of having just been struck (within a very short 

 time interval — see below) by a photon of the former frequency, it may be 

 absorbed and transfer the atom into state C. From this state the atom 

 may proceed spontaneously to state B, so that emission of the line BC 

 will be produced by simultaneous irradiation with the two frequencies 

 NA and AC, though not by irradiation with either frequency separately. 

 Now suppose that the intensity of both of the irradiating lines is doubled. 

 The intensity of the Hne BC emitted from the gas will then be quadrupled; 

 for the chance of a given atom being struck by a photon of the frequency 

 AC within a short time interval after it is struck by a photon of the 



