( 888 ) 
part of its energy. In that case we are already witnessing an “ab- 
sorption” process; the further transformation of the energy into 
heat, ete. ensues by collisions. 
According to this conception, absorption of radiant energy only 
takes place, when by resonance certain electrons are set vibrating to 
such a degree, that energy is imparted by them in an irreversible 
way to other parts of the systems to which they belong. As to the 
particular conditions of this process, we can only guess at them. 
Lorentz has shown?) that the damping influence of collistons may 
approximately be expressed by introducing into the equations of 
motion a term, proportional to the velocity of the electron. But if 
the absorption process already begins within the molecule, before a 
new encounter takes place (as Lorentz thinks probable, 1. ¢. p. 142), 
it is quite conceivable that the amplitude, and, therefore, the velocity 
of the resonant electron must have increased beyond a certain limit, 
before a continual transfer of energy to otber parts of the same 
molecule can result? If the connections really are of that kind, then 
waves which on account of imperfect resonance impart only small 
mean velocities to the electron, will suffer no absorption at all: the 
part of the damping parameter that is due to absorption, will sink 
to zero at a certain smal distance on both sides of the centre of the 
absorption line. This, of course, is an hypothesis to which, in the 
absence of a deeper knowledge of the internal structure of the mole- 
ecules, we are unable to give a solid foundation, but which may be 
put to the test by scrutinizing the deductions following from it. 
Suppose a beam of white light to pass through a rarefied gas 
having rather sharply defined periods of free vibrations *), then, 
according to § 1, the spectral region in which effectual co-vibrating 
oceurs, is only little wider — for each kind of electrons — than 
1) Lorentz, 1, c. Note 57. 
2) This conception agrees very well with a new radiation hypothesis, recently 
proposed by Prancx (Verhandl. d. Deutschen physikalischen Gesellschaft 13, p. 138 , 
according to which accumulating radiant energy by resonance is a continuous 
process, whereas emitting radiant energy only takes place by definite “light- 
quanta”. — Now, let the forced vibrations of an imperfectly resonant electron 
attain an intensity sufficient for it to emit light-quanta by itself, but not sufficient 
to shake the other electrons, belonging to the same molecule (and having frequen- 
cies of their own), to such a degree as would be necessary for them also to 
emit quanta; then part of the incident radiation is scattered by the electron first 
affected, but there is no absorption, no transformation of the accumulated energy 
into energy of some other kind. 
5) The proper periods are never defined with perfect sharpness, owing to the 
disturbing influence of collisions and to the Dorrrer-effect. 
