820 



SCIENCE. 



[N. S. Vol. XXIV. No. 625. 



in a particular form is imparted to a sub- 

 stance it spreads out very slowly among the 

 various possible modes of molecular motion 

 and if the substance is losing energy continu- 

 ously by radiation we must have a very wide 

 departure from black body radiation because 

 of the wide and persistent departure of the 

 substance from thermal equilibrium. I think 

 this departure of the radiation of a substance 

 from black body radiation should be attributed 

 to its actual cause, selective excitation, and it 

 should not be spoken of as selective radiation 

 in the strict sense of that term. 



The extent to which the radiation from a 

 selectively excited substance departs from 

 black body radiation or rather from its own 

 characteristic normal emission (when it is 

 nearly in thermal equilibrium) depends great- 

 ly upon the speed at which the energy of a 

 given mode of molecular motion spreads out 

 into all the possible modes, and we have evi- 

 dence that this speed of spreading is very 

 slow even in many solid and liquid substances. 

 Thus we have in the fire-fly a case of selective 

 excitation and the wide departure of the radi- 

 ation of the fire-fly from normal black body 

 radiation shows that the energy which is de- 

 veloped by the selective excitation is nearly 

 all radiated before it spreads out to any great 

 extent among the various possible modes of 

 motion. 



In the case of the Welsbach mantle it is not 

 at all certain that we have a genuine case of 

 selective radiation free from the effects of 

 selective excitation, for, although the exciting 

 agent in this case is the extremely disordered 

 movements of combustion, still even the dis- 

 ordered movements of combustion do no 

 doubt depart very widely from the type of 

 molecular motion which would exist in the 

 same substance in thermal equilibrium. 



When energy is imparted to a glowing sub- 

 stance by the electric current, whether the 

 substance be solid, or liquid, or gas, we have 

 in all probability a strongly marked case of 

 selective excitation. 



The upshot of this whole matter is that in 

 the solution of the important problem of the 

 efficient production of light we are not con- 

 strained by the thermodynamic laws of radia- 



tion, and not to a very great degree dependent 

 upon selective radiation properly so called, but 

 we are left free in the field of unlimited possi- 

 bilities of selective excitation and we may look 

 forward with some hope of a highly efficient 

 lamp independently of the discovery of a sub- 

 stance which will stand temperatures of many 

 thousands of degrees. 



In the tungsten lamp we have certainly a 

 filament which stands a very high temperature 

 (several hundred degrees higher than the car- 

 bon filament can stand), we have a filament 

 which certainly shows selective radiation in 

 the strict sense of this term, and we have most 

 certainly some degree of selective excitation. 

 To what extent the high efficiency of the 

 tungsten lamp is to be attributed to one or 

 another of these three things it is impossible 

 to decide from present data. 



In the mercury-vapor lamp and in the 

 titanium-are lamp we have certainly a sub- 

 stance (a vapor) which can stand an unlim- 

 ited degree of temperature, but we know that 

 the vapor is not very hot in either case; also 

 in both lamps the light-giving vapor most 

 certainly shows selective radiation in the strict 

 sense of this term, and in both cases we most 

 certainly have very pronounced selective ex- 

 citation. Furthermore, in the case of a gas 

 or vapor it seems that the speed of spreading 

 out of a given mode of molecular motion into 

 all possible modes is very slow, so that select- 

 ive excitation in a vapor or gas shows itself 

 in very pronounced departure of the radiation 

 from the normal characteristic radiation of 

 the given gas or vapor. 



The most striking instance of selective ex- 

 citation as shown by extremely abnormal 

 radiation is that afforded by the long-con- 

 tinued glow of the air in a Geissler tube at 

 liquid-air temperature after the cessation of 

 the exciting current. It seems, indeed, that 

 the lower the temperature of a substance the 

 slower the energy of a given exaggerated mode 

 of molecular motion spreads into other modes, 

 and the higher the temperature the more this 

 spreading is accelerated. A remarkable con- 

 sequence of which is that a selectively excited 

 gas should be cold to give the greatest possible 

 luminous efficiency, whereas a very hot gas 



